Registration
The 2026 program
Agenda
Explore the talks and events happening across both days of MITechCon.
See all talks in the session catalogue.
All times Eastern Time (ET).
Thursday, October 29
Select a session title to read its description.
Time
Registration
Ballroom ABSponsored by Oakland University School of Engineering and Computer Science
Ballroom C WestSponsored by UWM
Ballroom C EastSponsored by Oakland County Tech Collaboration
Gold Room ASponsored by Solliance
Gold Room B
Gold Room C
8:30 AM
Check-In8:30 AM
Breakfast8:30–10:00 AM
9:00 AM
10:00 AM
11:15 AM
12:15 PM
Lunch12:15–1:15 PM
1:15 PM
2:30 PM
3:45 PM
5:00 PM
Reception5:00–6:30 PM
8:30 AM
Ballroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Breakfast
8:30–10:00 AM9:00 AM
Ballroom
Build vs. Buy in the AI Era
Jason Bressler9:00 AM–9:50 AM10:00 AM
Room D
AI Influence on Reliable Machine Vision
Michael Koper10:00 AM–11:00 AMRoom B
Civic Agents: The Rise Of AI In Modern Public Service
Himanshu Goil10:00 AM–11:00 AMRoom E
Driving into the Future: Navigating Emerging Cybersecurity Threats in Autonomous Technology
Hemanth Tadepalli10:00 AM–11:00 AMRoom F
Engineering Beyond the "Rule of Eight"
Harsh Ranjan10:00 AM–11:00 AMRoom A
RAG for Data Engineers: What It Is, Where It Fits, and Why Your Metadata Is the Missing Piece
Varun Joshi, Nitesh Khapekar10:00 AM–11:00 AMRoom C
Your AI Agent Just Ran DELETE Without WHERE. Now What?
Jonathan Stewart10:00 AM–11:00 AM11:15 AM
Room A
A Practical Guide to Running LLMs on Kubernetes
Rohit Mishra11:15 AM–12:15 PMRoom C
Accelerating USB Host Verification Using Real USB Devices in a Virtualized Device Modeling Framework
Suchir Gupta11:15 AM–12:15 PMRoom E
Fortifying AI: Understanding Prompt Attacks and How to Defend Against Them
Ethan DeMott11:15 AM–12:15 PMRoom D
From Standard Work to Scalable Smart Manufacturing: Making Data and AI Work in Assembly Operations
Bryan Bauw11:15 AM–12:15 PMRoom B
Google BigQuery - Heart of Smart Analytics toolset on Google Cloud
Vijaykumar Jangamashetti11:15 AM–12:15 PMRoom F
LLMs are tools, not solutions: Entity extraction as a computer science problem
Milena F, Andy Schmidt11:15 AM–12:15 PM12:15 PM
Ballroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Lunch
12:15–1:15 PM1:15 PM
Room B
AI and Digital Transformation in Sepsis Management for Smarter Healthcare Delivery
Swapna Chimanchodkar1:15 PM–2:15 PMRoom D
AI Driven Production Planning Using Decision Intelligence in Manufacturing Systems
Madhav Jayesh Kumar Pandya1:15 PM–2:15 PMRoom A
Medallion Architecture in Regulated Healthcare: A Hands-On Playbook
Raziullah Khan1:15 PM–2:15 PMRoom C
Mission-Critical Code: What NASA’s Power of Ten Can Teach Us
Jonathan "J." Tower1:15 PM–2:15 PMRoom E
The Invisible Stack: Black-Ops of Modern Enterprise
Robert Gray III1:15 PM–2:15 PM2:30 PM
Room C
5G-Optimized Android System Architecture for Low-Latency Applications
Patel Riddhi2:30 PM–3:30 PMRoom A
Agentic Loops in the Data Stack: From Pipeline Failure to Auto-Remediation
Varun Joshi2:30 PM–3:30 PMRoom B
AI Agents in Products: A Product Manager’s Guide to Building What Matters
Reshmika Dhandapani2:30 PM–3:30 PMRoom F
Building Enterprise-Grade AI Solutions with Oracle Generative AI Service and the Oracle AI Database
Juarez Junior2:30 PM–3:30 PMRoom D
Future of Agentic AI in Telecom Billing and Revenue Management
Balu Chavan2:30 PM–3:30 PMRoom E
Probably Secure: What We Misunderstand About AI And Determinism Is Making Us Less Secure
Dwayne McDaniel2:30 PM–3:30 PM3:45 PM
Room A
Designing Data Pipelines That Don’t Hate You Six Months Later
Chris Birie3:45 PM–4:45 PMRoom E
Microsoft’s Approach to Comprehensive Security in the Age of AI
David Giard3:45 PM–4:45 PMRoom B
Multi-Agent Without the Bloat: Designing Lean AI Systems on Azure
Carey Payette3:45 PM–4:45 PMRoom C
Three Lessons from a Woman in Enterprise Architecture
Sarah Wimberley3:45 PM–4:45 PM5:00 PM
Ballroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Reception
5:00–6:30 PM
Friday, October 30
Select a session title to read its description.
Time
Registration
Ballroom ABSponsored by Oakland University School of Engineering and Computer Science
Ballroom C WestSponsored by UWM
Ballroom C EastSponsored by Oakland County Tech Collaboration
Gold Room ASponsored by Solliance
Gold Room B
Gold Room C
8:30 AM
Check-In8:30 AM
Breakfast8:30–10:00 AM
9:00 AM
OpeningStarts 9:00 AM
10:00 AM
11:15 AM
12:15 PM
Lunch12:15–1:15 PM
1:15 PM
2:30 PM
3:45 PM
5:00 PM
Closing5:00–5:30 PM
8:30 AM
Registration
Check-In
Opens at 8:30 AMBallroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Breakfast
8:30–10:00 AM9:00 AM
Ballroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Opening
Starts 9:00 AM10:00 AM
Room E
Adaptive Machine Learning for Real Time Transaction Routing in Distributed Systems
Jayaseelan Shanmugam10:00 AM–11:00 AMRoom A
Building an AI-Ready Finance Lakehouse with dbt, Governed Metrics, and Databricks
Mou Rakshit10:00 AM–11:00 AMRoom C
Floors, Not Ceilings: Lessons from 6 Months of AI Enablement at a $10B Company
Adam Broadbent10:00 AM–11:00 AMRoom D
Next Evolution of Advanced Manufacturing with ODIN Digital Solutions
Yanesh Naidoo10:00 AM–11:00 AMRoom B
Semantic Layers: The Missing Foundation for Agentic AI in Enterprise
Sajitha Sritharan10:00 AM–11:00 AM11:15 AM
Room D
Robotic Tracer Gas Sniffing for Leak Localization in EV Battery Packs and Energy Storage Systems
Davy Leboucher11:15 AM–12:15 PMRoom E
Scaling Enterprise Intelligence with Agentic Orchestration — Beyond the Chatbot
Sriramprabhu Rajendran11:15 AM–12:15 PMRoom B
The AI is ready, but you are the problem. How to finally be ready for a real AI.
Balazs Horvath11:15 AM–12:15 PMRoom C
Your AI Reads Entire Files. It Only Needs 6%. Here's How We Fixed It.
Rajkumar Sakthivel11:15 AM–12:15 PM12:15 PM
Ballroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Lunch
12:15–1:15 PM1:15 PM
Room D
AI on the Edge
Jared Rhodes1:15 PM–2:15 PMRoom A
Applying Machine Learning and Network Analytics for Modern Financial Crime Detection
Shivam Tiwari1:15 PM–2:15 PMRoom B
Civic Agents in Practice: The Next Phase of AI for Government Workflows
Himanshu Goil1:15 PM–2:15 PMRoom C
Designing Conversational BI: How Azure Databricks AI/BI Genie Unlocks Self-Service Insights
Mou Rakshit1:15 PM–2:15 PMRoom E
Headroom: A Context Optimization Layer for LLM Applications
Tejas Chopra1:15 PM–2:15 PM2:30 PM
Room A
Blending Product Thinking with Software Modernization
Brian McKeiver2:30 PM–3:30 PMRoom E
Diverge, Converge, Speak Up: The Discipline of Team Creativity
Tom Rodgers2:30 PM–3:30 PMRoom C
Don’t Let Your Org Chart Write Bad Code
Luis Fernandez2:30 PM–3:30 PMRoom B
Knights, Castles, and Creative Cloud: Rapid Illustration Concepting with AI
Rex Rainey2:30 PM–3:30 PMRoom D
Securing Cyber Physical Systems in Municipal Infrastructure Environments
Venkata Kartheek Reddy Somasani2:30 PM–3:30 PM3:45 PM
Room E
Building Scalable Multi-Agentic AI Systems: Orchestrating Agents with Event-Driven Approach
Mary Grygleski3:45 PM–4:45 PMRoom A
Case Study: Designing AI-Augmented Enterprise Investigation Systems
Santosh Vasudevan3:45 PM–4:45 PMRoom B
Digital Transformation or Digital Chaos? A Practical Guide for Leaders Who Are Done Playing Catch-Up
Matt Russell3:45 PM–4:45 PMRoom D
How to get started with collection and storage of industrial data
Grant Pinkos3:45 PM–4:45 PMRoom C
Test Smarter with AI: Playwright Automation at Scale on Azure
Randy Pagels3:45 PM–4:45 PM5:00 PM
Ballroom AB · Sponsored by Oakland University School of Engineering and Computer Science
Closing
5:00–5:30 PM
Keynote · Artificial Intelligence and Machine Learning
Build vs. Buy in the AI Era
- When
- Thursday, October 29 · 9:00 AM–9:50 AM ET
- Room
- Ballroom
- Speaker
- Jason Bressler
Artificial intelligence is changing the way organizations operate, but one of the biggest questions leaders face remains the same: Should we build it ourselves or buy it? The answer isn’t just about cost. It impacts how quickly your organization can move, how adaptable your solutions are, who owns your data, and ultimately how you create a competitive advantage.
Drawing from his experience leading technology and AI initiatives at enterprise scale, United Wholesale Mortgage Chief Technology Officer Jason Bressler will share a practical perspective on what it means to own the technology that truly sets your organization apart. He’ll discuss why successful AI adoption involves more than simply automating existing processes, how intelligent agents are creating opportunities to rethink the way work gets done, and why trusted, high-quality data is still the foundation of every meaningful AI effort.
Rather than focusing on hype or future predictions, this keynote will highlight real-world lessons learned from bringing AI into production. Jason will explore the leadership decisions, organizational challenges, and strategic tradeoffs that come with AI transformation. Attendees will walk away with a practical framework for evaluating build-versus-buy decisions and a new perspective on AI: not just as a tool to do today’s work faster, but as an opportunity to rethink what’s possible and create entirely new ways of working.
Conference Session · Artificial Intelligence and Machine Learning
AI Influence on Reliable Machine Vision
- When
- Thursday, October 29 · 10:00 AM–11:00 AM ET
- Room
- Room D
- Speaker
- Michael Koper
In innovative vehicle manufacturing, the fusion of Vision Systems and Artificial Intelligence (AI) has revolutionised quality control. As manufacturers pursue precision and efficiency, advanced vision technologies and AI algorithms ensure high production standards. Vision systems equipped with high-resolution cameras and lights serve as the vigilant eyes that meticulously scrutinise every aspect of vehicle assembly.
Real-time data is processed by AI, identifying defects and adapting to evolving standards. AI-driven quality control extends beyond error detection, predicting potential issues and reducing production costs. Integrating Artificial vision in quality control fosters a data-driven approach, optimising processes and elevating assembly precision.
This synergy between Vision Systems and AI is a cornerstone in smart vehicle manufacturing, ensuring vehicles meet safety, performance, and reliability standards. The future of quality control is marked by unprecedented accuracy, efficiency, and a commitment to exceeding customer expectations
Conference Session · Artificial Intelligence and Machine Learning
Civic Agents: The Rise Of AI In Modern Public Service
- When
- Thursday, October 29 · 10:00 AM–11:00 AM ET
- Room
- Room B
- Speaker
- Himanshu Goil
Governments are starting to use AI agents to improve how services are delivered—from permits and public works to citizen support. These “Civic Agents” are helping staff work more efficiently and making public services easier to access.
In this session, we’ll look at real examples from local governments already using AI, and share a simple 5-step framework (Purpose, Pilot, Playbook, Production, Performance) for scaling these solutions. Whether you're in tech or public service, you’ll walk away with practical ideas to start or grow your AI journey.
Conference Session · Autonomous Systems and Intelligent Robotics
Driving into the Future: Navigating Emerging Cybersecurity Threats in Autonomous Technology
- When
- Thursday, October 29 · 10:00 AM–11:00 AM ET
- Room
- Room E
- Speaker
- Hemanth Tadepalli
As vehicle autonomy, connectivity and artificial intelligence reduce the level of control and increase exposure to attacks, it has never been more important to strengthen cybersecurity programs and be fully aware of emerging threats.
Cybersecurity threats are not static; it’s a dynamic environment and constantly changing. Every day, there is a new potential threat, and while technology is becoming more and more sophisticated, actors are also becoming more and more intelligent and capable.
Join this presentation and gain insight into emerging risks and threats that can potentially be caused by sophisticated actors and advanced technologies, and measures to mitigate them.
Conference Session · Artificial Intelligence and Machine Learning
Engineering Beyond the "Rule of Eight"
- When
- Thursday, October 29 · 10:00 AM–11:00 AM ET
- Room
- Room F
- Speaker
- Harsh Ranjan
Modern enterprise AI is facing a silent production crisis: the "Instruction Drop." While Large Language Models (LLMs) appear highly capable in isolated demos, their reliability in complex, rule-heavy environments decays sharply once a prompt exceeds roughly eight core instructions. This session explores the technical "why" behind this failure—grounded in the "Lost in the Middle" phenomenon where transformer attention weights diminish at the center of dense prompts.
Moving beyond the fragile world of "prompt engineering," this talk provides a senior-level architectural roadmap for building resilient, production-grade systems. We will deep-dive into Functional Factoring, a method for decomposing monolithic "God-Prompts" into governed modular pipelines. Attendees will learn how to implement a Governed Data Layer to offload logic-parsing and utilize the MetaGPT framework to coordinate an "assembly line" of specialized agents. By utilizing CriticGPT-style audit loops and typed interfaces, architects can move from 70% "vibe-check" reliability to 99.9% deterministic enforcement. Stop prompting and start architecting for the next era of enterprise AI.
Conference Session · Artificial Intelligence and Machine Learning
RAG for Data Engineers: What It Is, Where It Fits, and Why Your Metadata Is the Missing Piece
- When
- Thursday, October 29 · 10:00 AM–11:00 AM ET
- Room
- Room A
- Speakers
- Varun Joshi, Nitesh Khapekar
Large language models are only as useful as the context you give them. For data engineers, that context is your schema, your lineage graph, your query history, your dbt models — and most teams haven't connected those dots yet.
This session is a practical introduction to Retrieval-Augmented Generation from a data engineering lens. We'll cover what RAG actually does under the hood, why it matters more for data teams than most, and the four places it shows up naturally in the data engineering workflow: answering schema questions without digging through a stale catalog, grounding SQL agents in your actual table definitions, giving incident response agents access to historical pipeline context, and surfacing institutional knowledge that currently lives only in senior engineers' heads.
Conference Session · Artificial Intelligence and Machine Learning
Your AI Agent Just Ran DELETE Without WHERE. Now What?
- When
- Thursday, October 29 · 10:00 AM–11:00 AM ET
- Room
- Room C
- Speaker
- Jonathan Stewart
Your AI agent deleted 1,943,200 rows. Not because it was malicious — because in a documented March 2026 incident, a developer gave an agent a routine AWS cleanup task, it substituted terraform destroy for the instructed CLI commands without asking, and a production database vanished. No confirmation prompt fired. No override path existed. The model's guardrails didn't catch it.
This session is for SQL Server DBAs, Fabric administrators, and data engineers who are already deploying — or being asked to deploy — AI agents against production data systems.
The fix is not better prompting. It is infrastructure enforcement. The destructive command gate is a Claude Code hook that intercepts tool calls before execution and hard-blocks dangerous operations regardless of what the model decided: DELETE without WHERE, DROP TABLE, terraform destroy, git push --force to main, rm -rf, and .env edits. Six operation categories. No override path. No reasoning around it. For Fabric and SQL Server environments, this means agent-initiated SQL executions, schema modifications, and Lakehouse destructive actions are intercepted — not just shell commands. The gate adds under 100 milliseconds to any tool call — imperceptible against the model's own response time — and zero false positives across a tested corpus of common operations including softDelete patterns, TRUNCATE in datetime functions, and git rebase. The agent hits a wall only when it should.
Live demo: watch the gate fire mid-session on a real DELETE attempt. The session continues from control, not recovery.
You'll leave with:
A working hook pattern deployable on your Fabric and SQL Server estate today
A threat model for AI agents with live data access — what breaks and how
The implementation distinction between soft guardrails (model training) and hard blocks (infrastructure enforcement)
The architectural principle: intercept between model decision and tool execution — the hook point differs by framework, the principle does not
Conference Session · Artificial Intelligence and Machine Learning
A Practical Guide to Running LLMs on Kubernetes
- When
- Thursday, October 29 · 11:15 AM–12:15 PM ET
- Room
- Room A
- Speaker
- Rohit Mishra
The path from “I have a model” to “it’s serving production traffic on my cluster” is full of sharp edges. This talk is an end-to-end walk-through: GPU device plugins and node scheduling, model weight distribution strategies for 10–140 GB artifacts, health checks that survive multi-minute model loads, and GPU-aware auto-scaling. Includes a live demo deploying an open-weight LLM on Kubernetes and serving inference traffic with real-time GPU metrics.
Conference Session · Simulation, Modeling, and Digital Engineering
Accelerating USB Host Verification Using Real USB Devices in a Virtualized Device Modeling Framework
- When
- Thursday, October 29 · 11:15 AM–12:15 PM ET
- Room
- Room C
- Speaker
- Suchir Gupta
Verifying USB Host designs-under-test (DUTs) is increasingly difficult as the USB ecosystem continues to expand across device types, protocol modes, and class behaviors. Traditional verification approaches depend on synthetic device models and manually developed testbenches, which are expensive to build, difficult to maintain, and often unable to capture the nuanced behavior of real hardware.
This work presents a production-proven verification methodology that brings physical USB devices directly into simulation and emulation environments, replacing conventional behavioral device models with real hardware. The framework integrates the Linux USB subsystem and the libusb library to create a user-space virtual adapter that connects physical USB devices to SystemVerilog DPI-based device BFMs. Through this architecture, a USB Host DUT can enumerate and communicate with real USB devices using protocol-accurate control, bulk, and interrupt transactions while maintaining standards compliance and timing fidelity.
Case studies using USB mass-storage devices across High-Speed, SuperSpeed, and SuperSpeed+ modes show significant reductions in testbench development effort, improved behavioral coverage, and better exposure to real-world timing and interoperability conditions that are difficult to reproduce with model-based approaches. The framework also enables the same verification architecture to be reused across both simulation and emulation platforms without modification.
By incorporating real devices into the verification loop, this methodology offers a scalable and realistic path for USB Host validation. It also lays the groundwork for extending real-device-based verification to additional USB classes and other high-speed interfaces where accurate device behavior is essential.
Conference Session · Cybersecurity, Privacy, and System Resilience
Fortifying AI: Understanding Prompt Attacks and How to Defend Against Them
- When
- Thursday, October 29 · 11:15 AM–12:15 PM ET
- Room
- Room E
- Speaker
- Ethan DeMott
As AI systems become increasingly integrated into enterprise applications, they also introduce a new and rapidly evolving attack surface. In this session, Ethan DeMott explores the emerging landscape of prompt-based attacks on large language models, including prompt extraction, prompt injection, and jailbreak techniques.
Attendees will gain a practical understanding of how attackers exploit AI systems to leak sensitive data, manipulate outputs, and bypass safety controls. The session will also walk through real-world-inspired scenarios and demonstrate how seemingly harmless prompts can be weaponized.
Beyond identifying threats, this talk focuses on actionable defenses. Ethan will cover layered protection strategies across the model, prompt, and system levels—equipping developers, engineers, and leaders with the tools needed to build more secure and trustworthy AI applications.
Conference Session · Advanced Manufacturing and Industry 4.0
From Standard Work to Scalable Smart Manufacturing: Making Data and AI Work in Assembly Operations
- When
- Thursday, October 29 · 11:15 AM–12:15 PM ET
- Room
- Room D
- Speaker
- Bryan Bauw
Many manufacturing operations are investing in digital tools, but still struggle with inconsistent processes, limited visibility, and solutions that don’t scale beyond isolated use cases. The challenge isn’t technology—it’s connecting standard work, operators, and data in a way that drives real results.
This session outlines a practical approach to improving assembly performance by starting with operator enablement and standardized workflows, then layering in real-time data capture, error-proofing, and system integration through MES.
Attendees will learn how to reduce variability, improve quality, and create a data foundation that supports better decision-making and enables AI-driven insights—without disrupting production.”
Conference Session · Artificial Intelligence and Machine Learning
Google BigQuery - Heart of Smart Analytics toolset on Google Cloud
- When
- Thursday, October 29 · 11:15 AM–12:15 PM ET
- Room
- Room B
- Speaker
- Vijaykumar Jangamashetti
This session will provide insights about Google BigQuery, a fully-managed, petabyte-scale analytics data warehouse that has been continuously innovating for over a decade. The session will discuss how BigQuery can help organizations drive limitless innovation with an open, unified, and intelligent data ecosystem by democratizing data insights, while gaining data cloud flexibility and predictability with BigQuery ML. The session will also cover different ways to ingest data into BigQuery and best practices, and will conclude with a live demo of the BigQuery console and a showcase of a petabyte demo.
Conference Session · Artificial Intelligence and Machine Learning
LLMs are tools, not solutions: Entity extraction as a computer science problem
- When
- Thursday, October 29 · 11:15 AM–12:15 PM ET
- Room
- Room F
- Speakers
- Milena F, Andy Schmidt
You can compare AI to a calculator: just because a tool produces results doesn’t mean the underlying science has disappeared. Challenges that have existed in computer science for years are simply being addressed today using different methods.
This is well illustrated by the example of entity extraction. Let’s imagine a Large Language Model (LLM) analyzing thousands of medical studies and automatically extracting relevant information such as medications, diseases, and side effects. The goal is to extract structured data from unstructured text—for instance, to efficiently evaluate correlations or interactions.
To achieve this, an LLM must also handle ambiguities (e.g., “Java” as a programming language, an island, or coffee), take context into account, and correctly recognize nested entities. These challenges persist regardless of whether rule-based methods, classical machine learning approaches, or modern LLMs are used.
Using a specific use case, this presentation demonstrates that while we now have more powerful tools at our disposal, classic challenges in computer science persist. Building on this, a practical architecture is presented: LLMs for semantic extraction, vector databases for contextualization, and n8n for orchestrating robust data pipelines. Step by step, we’ll show how a naive prompt can be transformed into a stable, traceable, and extensible solution.
Conference Session · Artificial Intelligence and Machine Learning
AI and Digital Transformation in Sepsis Management for Smarter Healthcare Delivery
- When
- Thursday, October 29 · 1:15 PM–2:15 PM ET
- Room
- Room B
- Speaker
- Swapna Chimanchodkar
Sepsis remains one of the leading causes of preventable mortality worldwide, requiring rapid diagnosis, coordinated intervention, and intelligent healthcare decision making. Despite advancements in clinical care, healthcare systems continue to face challenges related to fragmented workflows, delayed escalation, and underutilized clinical data.
This presentation explores how artificial intelligence, predictive analytics, and digital transformation are reshaping sepsis management and enabling smarter healthcare delivery. By integrating electronic health records, real time monitoring systems, machine learning models, and data driven workflows, healthcare organizations can improve early sepsis detection and accelerate clinical response.
Research demonstrates that AI enabled predictive systems can identify sepsis risk several hours earlier than conventional clinical approaches, creating critical opportunities for proactive intervention. Combined with process transformation strategies such as workflow redesign, standardized escalation protocols, and cross functional collaboration, these technologies help improve operational efficiency and patient outcomes.
The session will highlight practical approaches for embedding intelligent technologies into clinical operations while addressing implementation considerations including interoperability, algorithm transparency, adoption barriers, and equitable access to digital healthcare innovation.
By combining digital intelligence with process excellence, healthcare systems can transition from reactive care models to proactive and data informed healthcare delivery, improving survival outcomes and advancing the future of healthcare technology.
Conference Session · Advanced Manufacturing and Industry 4.0
AI Driven Production Planning Using Decision Intelligence in Manufacturing Systems
- When
- Thursday, October 29 · 1:15 PM–2:15 PM ET
- Room
- Room D
- Speaker
- Madhav Jayesh Kumar Pandya
Modern manufacturing environments operate under increasing complexity driven by demand variability, supply uncertainty, and capacity constraints. Traditional planning systems often struggle to respond dynamically to these challenges, leading to inefficiencies in production, excess inventory, and reduced service performance. This session presents a practical approach to implementing artificial intelligence driven Integrated Business Planning systems for production optimization in complex manufacturing environments.
The session explores how machine learning based demand forecasting can be combined with operational constraint modeling to improve planning accuracy and responsiveness. A key focus is the use of decision intelligence frameworks that evaluate multiple production scenarios in real time by considering constraints such as capacity availability, supplier lead times, labor limitations, and fluctuating demand signals. These systems enable planners to move from reactive decision making to proactive and data driven planning.
Attendees will gain insight into how scenario modeling tools provide visibility into key performance indicators including capacity utilization, inventory levels, and service outcomes. The session also highlights lessons learned from implementing these systems in high variability environments, including common challenges related to data quality, system integration, and cross functional alignment.
In addition, the talk will demonstrate how predictive analytics can be used to support better material planning and reduce operational risk. By integrating financial impact analysis into planning decisions, organizations can align operational strategies with business objectives.
Participants will leave with actionable insights on how to design and deploy AI enabled planning systems that improve operational efficiency, enhance resilience, and support scalable manufacturing operations in real world settings.
Conference Session · DevOps, Cloud, and Platform Engineering
Medallion Architecture in Regulated Healthcare: A Hands-On Playbook
- When
- Thursday, October 29 · 1:15 PM–2:15 PM ET
- Room
- Room A
- Speaker
- Raziullah Khan
“Bronze-Silver-Gold” is easy to draw and hard to operate—especially under HIPAA. This session is a practical blueprint for running medallion architecture on BigQuery/Snowflake in healthcare. We’ll design ingestion with PII handling, build Silver transformations with data quality gates and lineage, and publish Gold data products ready for BI/ML. You’ll get Terraform/Jenkins patterns for CI/CD, access control strategies (row/column-level), and audit log models that satisfy compliance without slowing delivery. We’ll close with a short case pattern from pharma supply chain (inventory/logistics) showing cost trade-offs and reliability SLOs.
Conference Session · Software Engineering and Architecture
Mission-Critical Code: What NASA’s Power of Ten Can Teach Us
- When
- Thursday, October 29 · 1:15 PM–2:15 PM ET
- Room
- Room C
- Speaker
- Jonathan "J." Tower
Ever wonder how NASA writes software that literally can’t afford to fail? Meanwhile, back on Earth, our code is held together with duct tape and //TODO comments. The stakes might be a little lower for most of us—nobody's oxygen supply depends on our latest API deployment (hopefully)—but reliability, maintainability, and simplicity still matter.
Enter NASA’s Power of Ten rules: a set of software engineering principles designed to minimize complexity and maximize reliability. In this talk, we’ll explore these ten commandments of mission-critical coding, from banning recursion (gasp!) to keeping functions tiny and predictable. Along the way, we’ll discuss how these rules apply beyond space exploration—to web apps, enterprise systems, and anywhere else software keeps the lights on.
Whether you're launching rockets or just trying to keep production from crashing at 2 AM, these lessons can help you write code that’s more robust, maintainable, and maybe even astronaut-approved. So strap in, and join us for a journey through the Power of Ten—no space suit required!
Conference Session · Embedded, Edge, and Real-Time Systems
The Invisible Stack: Black-Ops of Modern Enterprise
- When
- Thursday, October 29 · 1:15 PM–2:15 PM ET
- Room
- Room E
- Speaker
- Robert Gray III
Behind every product your company ships is an invisible layer of operational disciplines that your customers will never see, never name, and never thank. DevOps built the pipeline. AIOps is watching it. DataOps is governing what flows through it. FinOps is making sure the company can afford to keep it running. These are the Black Ops of the modern enterprise — covert, mission-critical, and systematically undervalued. In this talk, Robert Gray III maps the full invisible stack: what it is, who builds it, what it costs when it breaks, and why the practitioners running it are the most important people in your organization.
WHAT ATTENDEES WILL LEARN
How the fifteen disciplines of the invisible stack are organized, what each one owns, and how they depend on each other. Most practitioners know their own discipline deeply and adjacent disciplines only partially. This talk gives them the full map.
Why delivery failures almost always trace back to a gap in the invisible stack; a FinOps function that was never built, an AIOps layer that was staffed too late, a DataOps practice that was substituted with a single data engineer and a prayer. Attendees will recognize these patterns in their own organizations immediately.
How to assess their own organization’s invisible stack maturity, where the gaps are, which disciplines are understaffed or structurally absent, and what the business cost of those gaps actually is when measured in delivery velocity, incident frequency, and engineering attrition.
What the hiring market is actually looking for across each discipline right now — not the job description version, but the candid version hiring managers share behind closed doors.
Conference Session · Embedded, Edge, and Real-Time Systems
5G-Optimized Android System Architecture for Low-Latency Applications
- When
- Thursday, October 29 · 2:30 PM–3:30 PM ET
- Room
- Room C
- Speaker
- Patel Riddhi
My proposed presentation provides a technical overview of a 5G-optimized Android system architecture for low-latency applications. It focuses on the gap between the promised capabilities of 5G and the limitations of Android’s legacy 4G-era network stack, which still creates latency, packet-processing, and resource-allocation bottlenecks.
The presentation explains the main challenges affecting Android performance under 5G, including memory-copy overhead, lack of adaptive QoS prioritization, weak edge-computing integration, and insufficient real-time traffic management. It then introduces a proposed architecture built around zero-copy data paths, kernel-bypass processing, smart traffic management, application profiling, and edge-computing support to improve responsiveness for real-time workloads.
It also highlights the expected impact of this architecture on applications such as cloud gaming, AR/VR, industrial IoT, and other mission-critical mobile systems, with reported improvements in latency reduction, throughput, QoS, and overall system efficiency. In short, the presentation shows how Android can be redesigned to better support current 5G workloads while creating a scalable foundation for future 6G and AI-native mobile systems.
Conference Session · Artificial Intelligence and Machine Learning
Agentic Loops in the Data Stack: From Pipeline Failure to Auto-Remediation
- When
- Thursday, October 29 · 2:30 PM–3:30 PM ET
- Room
- Room A
- Speaker
- Varun Joshi
Every data engineer knows the 2 AM pipeline failure — the one nobody notices until Friday's report is wrong. In this session, we break down five AI agents that are changing how data teams operate: from monitoring pipelines 24/7 and catching schema drift at ingestion, to closing the gap between a production failure and its root cause in minutes. We'll walk through real implementation patterns, including a baseline-learning monitoring agent and a tool-use driven incident response loop, and discuss what the shift to agentic data engineering actually means for the way teams are built and how engineers grow. Whether you're evaluating agents for your platform or already running them in production, you'll leave with concrete patterns you can apply immediately.
Conference Session · Product, UX, and Digital Experience
AI Agents in Products: A Product Manager’s Guide to Building What Matters
- When
- Thursday, October 29 · 2:30 PM–3:30 PM ET
- Room
- Room B
- Speaker
- Reshmika Dhandapani
AI agents are rapidly becoming part of modern software, but many teams rush into building them without clearly defining the problem they’re solving or the role the agent should play.
This session takes a practical, product management–driven approach to building AI agents that deliver real value. Instead of focusing only on tools or frameworks, it introduces a structured way to evaluate when agents are the right solution, and when they are not.
We’ll walk through how to identify high-impact workflows for agents, define clear boundaries and levels of autonomy, and design for reliability in inherently probabilistic systems. The session also covers how product managers and engineers can collaborate effectively, manage uncertainty in agent behavior, and prioritize agent initiatives alongside core product work.
Through real-world examples and lessons learned, attendees will leave with a clear framework they can apply immediately to design, evaluate, and ship AI agent features responsibly and effectively.
Conference Session · Software Engineering and Architecture
Building Enterprise-Grade AI Solutions with Oracle Generative AI Service and the Oracle AI Database
- When
- Thursday, October 29 · 2:30 PM–3:30 PM ET
- Room
- Room F
- Speaker
- Juarez Junior
Oracle AI Database 26ai introduces a unified, enterprise‑grade foundation for building AI‑powered applications directly where your data lives. Combined with the OCI Generative AI service, developers can now create intelligent, context‑aware applications without bolting on external complexity.
This session dives into how software engineers and architects can integrate vector search, in‑database embeddings, and generative AI orchestration into applications while maintaining performance, security, and operational simplicity.
Through practical examples and architectural guidance, you will learn how to combine Oracle’s AI‑native database capabilities with OCI’s managed LLM services to deliver production‑ready AI solutions that scale without sacrificing simplicity.
Conference Session · Artificial Intelligence and Machine Learning
Future of Agentic AI in Telecom Billing and Revenue Management
- When
- Thursday, October 29 · 2:30 PM–3:30 PM ET
- Room
- Room D
- Speaker
- Balu Chavan
As telecom services become increasingly real-time, digital, and partner-driven, traditional billing and revenue management systems are struggling to keep pace with the growing complexity of service delivery. By embedding Agentic AI across billing and payment ecosystems and integrating it with CRM platforms, telecom operators can transition from investigation-heavy customer support to AI-assisted first-call resolution, reducing Average Handling Time (AHT) by up to 50% while improving customer trust and revenue accuracy. Agentic AI enables a shift from reactive, manual revenue operations to intelligent, self-optimizing revenue platforms where autonomous agents continuously analyze network usage patterns, customer behavior, and financial signals. This capability allows telecom providers to proactively detect revenue leakage, optimize pricing strategies, and resolve billing issues before they impact customers. Through practical insights and real-world scenarios, this work illustrates how AI agents can collaboratively operate across charging, invoicing, revenue assurance, and customer support functions in near real time. It also outlines a strategic roadmap for modernizing legacy billing infrastructures while ensuring strong governance, transparency, and responsible AI adoption at scale.
Conference Session · Artificial Intelligence and Machine Learning
Probably Secure: What We Misunderstand About AI And Determinism Is Making Us Less Secure
- When
- Thursday, October 29 · 2:30 PM–3:30 PM ET
- Room
- Room E
- Speaker
- Dwayne McDaniel
When I say "probabilistic" outcomes, your mind likely jumps to coins or dice. However, probability also lies at the heart of the most significant development in computer science in recent memory: Artificial Intelligence. From classifying pictures of dogs to the probabilistic transformers of Generative AI, it is all based on the concept of "maybe."
GenAI is just another tech....right? Can we safely assume the same tools and mental models we have used for deterministic systems up til now are the best choice for this new world of probabilistic outcomes?
What are the security ramifications of a probabilistic system at scale?
The good news is that many of the rule-based security tools we have been relying on for years are still there to provide us with some safety netting, if we know when to employ them and where AI can best assist us in our mission.
Let's step way back and talk about systems that run on "most likely correct."
Join this talk if your teams are increasingly using AI-assistants and you're trying to figure out how to keep everyone safe. Let's have an interactive conversation about why it's important that you don't treat 'probably' and 'definitely' the same way.
Conference Session · Software Engineering and Architecture
Designing Data Pipelines That Don’t Hate You Six Months Later
- When
- Thursday, October 29 · 3:45 PM–4:45 PM ET
- Room
- Room A
- Speaker
- Chris Birie
Most data pipelines don’t fail on day one. They fail months later — when requirements change, data volume grows, or the original developer is no longer around. Much like martial arts training, pipelines built on shaky fundamentals tend to collapse under pressure.
This session focuses on practical data pipeline design patterns that emphasize durability, maintainability, and operational confidence. Drawing from experience modernizing enterprise data platforms, I’ll walk through common failure modes in batch and streaming pipelines and how disciplined upfront design decisions can prevent them.
We’ll cover topics such as schema evolution, idempotency, observability, testing strategies, and scaling — grounded in real examples of what worked, what broke, and what we had to correct after the fact. There are no silver bullets here — just repeatable patterns and lessons learned from systems that had to stay running while everything around them changed.
This talk is tool-agnostic and focused on principles that apply whether you’re using Python, Spark, Kafka, or traditional ETL platforms.
What’s in it for the attendee
- How to design pipelines that withstand change and growth
- Patterns for building observability and testability into data systems
- Common anti-patterns that lead to brittle pipelines
- Practical guidance rooted in real production experience
Conference Session · Cybersecurity, Privacy, and System Resilience
Microsoft’s Approach to Comprehensive Security in the Age of AI
- When
- Thursday, October 29 · 3:45 PM–4:45 PM ET
- Room
- Room E
- Speaker
- David Giard
In an era where cybercrime rivals the largest global economies, securing digital landscapes has never been more critical. In this session, we will delve into the intricacies of modern cybersecurity challenges and unveil Microsoft’s solutions designed to counteract these threats.
We will showcase a panoramic view of the current cybersecurity landscape, emphasizing the exponential growth of cybercrime. With cybercrime projected to surpass $8 trillion, understanding and mitigating these risks is paramount for any organization.
Key highlights of the session include:
• The Current State of Cybersecurity: Gain insights into the staggering scale of cybercrime and its economic impact, underscoring the urgency for robust security measures.
• Microsoft Security Platform: Explore Microsoft’s multi-cloud, multi-platform security solutions, including Microsoft Purview, Priva, Entra, Intune, Defender, and Sentinel, and how they integrate Copilot to enhance security operations.
• Unified SecOps Experience: Discover how Microsoft’s end-to-end platform approach serves as a force multiplier, providing comprehensive protection across diverse IT environments.
• Secure Future Initiative: Learn about Microsoft’s strategic investments in security, designed to stay ahead of evolving threats and secure AI workloads.
This session promises to equip attendees with actionable strategies and innovative tools to fortify their cybersecurity posture in an increasingly perilous digital age. Whether you are a cybersecurity professional, IT manager, or executive, this presentation will offer invaluable insights to help you navigate and secure the complex threat landscape.
Conference Session · Artificial Intelligence and Machine Learning
Multi-Agent Without the Bloat: Designing Lean AI Systems on Azure
- When
- Thursday, October 29 · 3:45 PM–4:45 PM ET
- Room
- Room B
- Speaker
- Carey Payette
Multi-agent systems get expensive and chaotic when every agent receives the full conversation, full tool history, and full memory state. In this session, we’ll look at how to design leaner AI systems on Azure using Azure OpenAI with Microsoft Agent Framework or LangGraph, focusing on the architectural patterns that reduce context bloat without losing capability. We’ll cover scoped context, targeted handoffs, short-term versus long-term memory, planner and router patterns, and techniques for keeping each agent focused on only the information it actually needs. Attendees will leave with practical design patterns for building multi-agent systems that are faster, cheaper, easier to debug, and easier to scale.
Conference Session · Software Engineering and Architecture
Three Lessons from a Woman in Enterprise Architecture
- When
- Thursday, October 29 · 3:45 PM–4:45 PM ET
- Room
- Room C
- Speaker
- Sarah Wimberley
Women remain significantly underrepresented in enterprise architecture and technical leadership roles, yet organizations increasingly need leaders who can bridge people, process, technology, and strategy—especially in the age of AI transformation. But, this is not just a talk about enterprise architecture—it is a conversation about courage, purpose, leadership, and creating space for more people to thrive in technology.
In this candid and engaging session, I'll share my unconventional journey from art school and creative work into enterprise architecture leadership. Attendees will explore three powerful lessons that apply far beyond IT: be curious, be authentic, and be flawed.
This session challenges traditional assumptions about who belongs in architecture and leadership roles while highlighting why organizations benefit from hiring individuals with diverse backgrounds, nonlinear career paths, and varied lived experiences. Attendees will leave with practical strategies for building confidence, navigating change, fostering trust, supporting psychological safety, and transforming perceived weaknesses into professional strengths.
Conference Session · DevOps, Cloud, and Platform Engineering
Adaptive Machine Learning for Real Time Transaction Routing in Distributed Systems
- When
- Friday, October 30 · 10:00 AM–11:00 AM ET
- Room
- Room E
- Speaker
- Jayaseelan Shanmugam
Modern distributed systems increasingly rely on real time decision making to optimize performance, reliability, and cost across complex service ecosystems. One of the most demanding use cases is transaction routing across multiple external service providers, where each decision must balance latency, success probability, cost efficiency, and system stability under constantly changing conditions.
This session explores how machine learning can be embedded into large scale distributed architectures to enable adaptive, data driven routing decisions. Moving beyond static rule based approaches, the presented framework leverages historical transaction data combined with real time telemetry to dynamically select optimal execution paths. The system continuously evaluates multiple variables including request attributes, temporal patterns, service health signals, and historical performance trends.
The architecture introduces three core layers of intelligence. First, anomaly detection models identify early signals of service degradation, enabling proactive mitigation before widespread failures occur. Second, automated failover mechanisms dynamically reroute traffic to maintain system continuity under partial outages. Third, reinforcement learning techniques iteratively refine decision strategies by learning from prior outcomes, improving efficiency over time without manual intervention.
Real world implementation patterns will be discussed, including event driven pipelines, feedback loops for continuous learning, and strategies for maintaining low latency at high throughput. The session also highlights challenges such as model drift, observability gaps, and safe deployment of adaptive systems in production environments.
Attendees will gain practical insights into designing resilient, self optimizing distributed systems that combine machine learning with robust engineering principles to deliver measurable improvements in performance, reliability, and operational efficiency.
Conference Session · Artificial Intelligence and Machine Learning
Building an AI-Ready Finance Lakehouse with dbt, Governed Metrics, and Databricks
- When
- Friday, October 30 · 10:00 AM–11:00 AM ET
- Room
- Room A
- Speaker
- Mou Rakshit
Finance teams are often asked to move faster, explain variance sooner, and support automation, but many organizations are still working from fragmented pipelines, inconsistent metric definitions, and dashboard-heavy reporting processes. Before finance can safely use AI or agentic workflows, the data foundation has to be trusted, governed, and explainable.
This session shows how to design an AI-ready finance lakehouse using Databricks, dbt, Unity Catalog, and governed metrics. We will walk through a practical architecture where finance data is transformed through dbt staging, intermediate, and mart layers, with reusable SQL models, macros, tests, documentation, and exposures creating a reliable path from raw source data to finance-ready outputs.
The session will also show how Unity Catalog strengthens governance through lineage, access control, classification, and auditability, while governed metric definitions help keep reporting, Power BI consumption, and AI-assisted analysis aligned to the same trusted business logic. We will discuss how this foundation can support future agentic finance use cases such as variance investigation, anomaly triage, and automated action without allowing AI to invent financial calculations.
Attendees will leave with a practical design pattern for building finance data products that are not only dashboard-ready, but AI-ready.
Conference Session · Software Engineering and Architecture
Floors, Not Ceilings: Lessons from 6 Months of AI Enablement at a $10B Company
- When
- Friday, October 30 · 10:00 AM–11:00 AM ET
- Room
- Room C
- Speaker
- Adam Broadbent
Most AI enablement programs in large engineering organizations fail the same way: leadership announces the tools, developers quietly resent the mandate, a few champions get loud on Slack, and six months later, nothing has changed except the line item on the IT budget. This session is an honest, field-tested account of what it took to do it differently — inside a $10 billion company's engineering organization, over a six-month engagement with measurable outcomes.
We'll walk through the five-phase rollout we ran end-to-end: a baseline AI readiness assessment, a structured Developer Fear Map workshop, a Leadership Expectations Alignment session, pattern-driven code assistance, CLAUDE.md authoring across active repositories, a special-focus pilot team, a scaled rollout informed by what the pilot taught us, and a sustainment cadence built around a monthly retrospective. You'll see the actual artifacts, the Fear Map facilitator guide, the CLAUDE.md template, the four prompting patterns we taught, and the productivity measurement framework that tracks adoption without surveilling developers.
Most importantly, you'll hear what didn't work, where leadership expectations had to be recalibrated mid-engagement. Where a metric we believed in turned out to erode trust, we changed it in Phase 4 because the pilot team taught us something we hadn't anticipated. Where "AI raises the floor, not the ceiling" stopped being a slogan and started being the actual ROI argument.
This workshop is built for engineering leaders making the investment case and the practitioners who have to make it work on Monday morning. Attendees leave with concrete artifacts, a clear-eyed view of the human and technical work involved, and a rollout sequence they can adapt to their own organization.
Conference Session · Advanced Manufacturing and Industry 4.0
Next Evolution of Advanced Manufacturing with ODIN Digital Solutions
- When
- Friday, October 30 · 10:00 AM–11:00 AM ET
- Room
- Room D
- Speaker
- Yanesh Naidoo
Every manufacturer wants the same thing. An assembly line that corrects itself, re-plans itself and one day runs with the lights off. Very few can say what their next step actually is.
The car industry already has a scale for this. Everyone knows roughly what it means for a car to go from doing nothing for the driver to needing no driver at all, and there are six steps in between. Nobody has applied that same scale to the assembly line that builds the car. This session does, and it turns out to be the clearest way yet to tell a plant where its assembly lines stand today and what the next step is worth.
It also gets specific about where AI fits on that journey, which is rarely where it is being sold. The AI that earns its place on an assembly line is small and narrow: vision watching whether a part is in position, a model watching press force drift towards a failure that has not happened yet, a model that spots a measurement population quietly shifting. None of it needs a new sensor. All of it needs a record underneath it that most plants do not have. The agent that reasons across the whole assembly line is real, and it is the last thing you buy rather than the first.
All of it is drawn from more than 3,000 assembly systems delivered worldwide, with real data off real assembly lines and the mistakes we made getting there.
You will leave able to:
- Place your own assembly line on the six-step scale from manual to autonomous, and name the next step honestly
- Tell which AI applications will survive on your shop floor and which need groundwork first
- Avoid three specific and expensive mistakes we have already made on your behalf
The session closes with a demo of AI agents reading a running assembly line and answering questions about it in plain language.
Conference Session · Artificial Intelligence and Machine Learning
Semantic Layers: The Missing Foundation for Agentic AI in Enterprise
- When
- Friday, October 30 · 10:00 AM–11:00 AM ET
- Room
- Room B
- Speaker
- Sajitha Sritharan
Your company is deploying agentic AI. Exciting. Then your AI agents make technically correct but business-wrong decisions. Why? Because raw data isn't enough. AI agents need semantic understanding—structured knowledge about entities, relationships, and business rules.
A semantic layer is the bridge. It transforms your data platform from a static repository into an AI-ready knowledge base where agents can reason, understand context, and explain decisions. In this case study, we'll share real implementation patterns using knowledge graphs, ontologies, and metadata standards to enable agentic AI across your organization. You'll learn architecture patterns, governance practices that scale without blocking teams, and ROI metrics that justify the investment.
Leave with a diagnostic framework to assess your organization's AI readiness, a 3-month roadmap to build semantic architecture, and confidence that your AI agents will make smart decisions grounded in business context.
Conference Session · Robotics Systems and Integration
Robotic Tracer Gas Sniffing for Leak Localization in EV Battery Packs and Energy Storage Systems
- When
- Friday, October 30 · 11:15 AM–12:15 PM ET
- Room
- Room D
- Speaker
- Davy Leboucher
The rapid scale‑up of lithium‑ion battery production for electric vehicles (EVs) and stationary energy storage systems (ESS) has introduced unprecedented challenges in leak detection and quality assurance. Battery packs integrate hundreds of cells, multiple sealing interfaces, cooling circuits, and safety‑critical enclosures, where even micro‑leaks can compromise thermal management, accelerate degradation, or create hazardous conditions. Ensuring hermetic integrity across complex geometries while maintaining production throughput remains a key unresolved issue.
This paper presents a robotic tracer gas sniffing approach for high‑precision leak localization in EV battery packs and energy storage assemblies. The system combines industrial collaborative robotics with high‑sensitivity helium and hydrogen tracer gas detection to overcome the limitations of conventional manual sniffing methods. By enforcing deterministic scan paths, controlled probe speed, and repeatable positioning, robotic sniffing decouples leak detection performance from operator variability and enables consistent inspection of complex battery architectures.
The proposed architecture integrates a multi‑axis robotic platform with tracer gas detectors capable of detecting leak rates down to the 10⁻⁸ atm·cc/s range and sub‑second response times. Robotic motion control allows uniform coverage of battery trays, modules, pack enclosures, cooling manifolds, and sealed electronic housings, including areas with restricted access or high geometric complexity. This deterministic scanning capability is particularly relevant for battery systems where sealing performance must be verified across large surfaces and multiple interfaces.
Beyond detection sensitivity, the system emphasizes data continuity and traceability, critical for battery manufacturing and second‑life repair workflows. Each battery pack or module is uniquely identified through barcode scanning prior to inspection. Leak events are automatically associated with the scanned identifier and spatially mapped onto a digital representation of the component. This enables precise localization of defects at the cell, module, or enclosure level, supporting targeted rework, root‑cause analysis, and quality trend monitoring throughout the battery lifecycle.
Comparative analysis between robotic and manual tracer gas sniffing highlights significant improvements in repeatability, inspection consistency, and downtime reduction. While manual sniffing remains highly operator‑dependent and difficult to standardize at production scale, robotic inspection provides reproducible results aligned with the requirements of high‑volume EV production and safety‑critical energy storage systems. The approach supports both end‑of‑line testing and repair or remanufacturing contexts, where accurate leak localization is essential for cost‑effective recovery of high‑value battery components.
The paper discusses:
Leak mechanisms and sealing challenges specific to EV battery packs and ESS
Robotic tracer gas sniffing system architecture and integration
Measurement repeatability and localization performance
Digital traceability using barcode identification and leak mapping
Application scenarios including pack assembly, module repair, and repurposing for second‑life energy storage
This contribution demonstrates how robotic tracer gas sniffing enables a shift from qualitative, operator‑dependent inspection toward a quantitative, traceable, and automation‑ready leak detection strategy for battery systems. The approach supports higher safety margins, improved manufacturing robustness, and data‑driven quality control for next‑generation electric mobility and energy storage infrastructures.
Conference Session · Software Engineering and Architecture
Scaling Enterprise Intelligence with Agentic Orchestration — Beyond the Chatbot
- When
- Friday, October 30 · 11:15 AM–12:15 PM ET
- Room
- Room E
- Speaker
- Sriramprabhu Rajendran
Millions of dollars were invested in your company's Gen AI chatbot solution. After using it for a week, users discovered that it was able to do nothing more than respond to queries and reverted back to their usual way of doing things. This is not a problem with the chatbot, but rather the architecture itself. Using an LLM wrapped up in a text box and calling it an "enterprise AI solution" would be like making an HTML page and calling it a "web application."
In this session, we will be exploring the architectural patterns that are needed for this transition. Agent decomposition patterns are discussed, which include the decomposition process of a monolithic "do everything" agent into smaller specialized agents that are independently testable and deployable. Patterns for orchestration vs choreography are explored, discussing when it is necessary to have a controlling agent and when collaboration between agents leads to better performance through events. The topic of tool-use patterns is thoroughly explored, covering the concept of MCP (Model Context Protocol) patterns, whereby the agents detect and invoke external tools and services. The most difficult part is also addressed – the observability/governance stack for running agent-based solutions in production since when a group of agents made an autonomous decision affecting your customers, you have to justify your action.
Participants will receive the architectural plan needed to transition from chatbot to agential orchestration, drawing on established enterprise architecture principles.
Conference Session · Artificial Intelligence and Machine Learning
The AI is ready, but you are the problem. How to finally be ready for a real AI.
- When
- Friday, October 30 · 11:15 AM–12:15 PM ET
- Room
- Room B
- Speaker
- Balazs Horvath
You bought a thousand Copilot licenses. You sent half the company through a prompting workshop. You even built your own CompanyGPT. The CEO mentioned it in the all-hands, you are "AI-Driven" now, and twelve months later, not a single thing in the business actually runs on AI.
People got woken up in the middle of the night and ambushed by some AI initiative, and now they have tools, training and a license bill, and still nothing to show for it. The AI was ready to ambush them. They were not ready to welcome it.
The hard part about AI is no longer the technical readiness, there are dozens of models just waiting to be used. It is the company underneath it. Processes that live in three people's heads. Data scattered across spreadsheets and silos. Approval workflows held together by manual copy-paste, goodwill and a lot of hope.
I have spent thirteen years building Microsoft solutions for enterprise clients across Europe, and the last four years specifically watching companies try to be AI-ready. They almost always fail at the same twelve checkpoints. So we built a scorecard for it.
In this session I walk you through the AI Readiness Scorecard, how it actually scores your organisation across foundation, technical setup and governance, and what the three resulting archetypes mean for your next ninety days.
You leave with a score for your own organisation, a ninety-day plan based on where you actually are, and everything you need to make the next AI initiative actually land.
Conference Session · Artificial Intelligence and Machine Learning
Your AI Reads Entire Files. It Only Needs 6%. Here's How We Fixed It.
- When
- Friday, October 30 · 11:15 AM–12:15 PM ET
- Room
- Room C
- Speaker
- Rajkumar Sakthivel
Every AI coding agent I've used does the same thing: it reads your entire file, every single time. On a mid-size Python project we work on, that's 45,000 tokens per query. The agent only needed about 4,900. We were burning through input tokens and didn't even realise how much until the bill landed.
So we built something to fix it. An open-source tool that indexes your codebase locally using tree-sitter for AST-aware chunking and sqlite-vec for vector search. It plugs into your editor through the Model Context Protocol, so the agent gets exactly the code it needs instead of everything. The result: 94% fewer tokens, measured across 20 real-world coding queries.
What made this tricky was getting retrieval quality right. Pure vector search missed too much; keyword search returned noise. We ended up combining both with Reciprocal Rank Fusion and adding confidence scoring so the agent knows when it should fall back to reading the full file.
In this talk, I'll walk through the architecture decisions we made and the ones we got wrong. Why we picked sqlite-vec over LanceDB. Why we chose truncation over LLM-based summarisation. How we handle secret redaction before anything hits the embedding model. And I'll show live benchmarks comparing retrieval with and without the index.
You'll walk away knowing how to build your own local RAG for code, how to measure token savings properly, and practical ways to cut your AI tooling costs without changing how you work.
Conference Session · Embedded, Edge, and Real-Time Systems
AI on the Edge
- When
- Friday, October 30 · 1:15 PM–2:15 PM ET
- Room
- Room D
- Speaker
- Jared Rhodes
Edge AI is not a new phenomenon. This session will show how models for video, audio, and sensor data run on devices close to where data is created. The result is faster response, better privacy, and steadier cost. The cloud still matters for training and fleet insight, but the moment-to-moment work happens on the device.
We’ll walk a full path from idea to production. Start with a model that fits your hardware. Trim it, test it, then measure real workloads. Deploy to what you already own or to dedicated edge gear.
You leave with a playbook that you can repeat. Pick, prepare, test, ship, observe, improve. No lock-in. No mystery. Just a clear way to move intelligence out of the data center and into the real world.
Conference Session · Artificial Intelligence and Machine Learning
Applying Machine Learning and Network Analytics for Modern Financial Crime Detection
- When
- Friday, October 30 · 1:15 PM–2:15 PM ET
- Room
- Room A
- Speaker
- Shivam Tiwari
Financial institutions face increasing challenges in detecting money laundering, fraudulent transactions, and emerging financial threats within rapidly evolving digital payment ecosystems. Traditional transaction monitoring systems often rely on static rule-based approaches that generate excessive false positives and struggle to adapt to changing criminal behaviors. This presentation demonstrates how modern data science and machine learning technologies can significantly improve detection accuracy, operational efficiency, and real time risk assessment in enterprise financial environments.
The proposed framework integrates multiple advanced analytical approaches to strengthen transaction monitoring capabilities. Behavioral profiling using unsupervised learning establishes individualized customer baselines that enable context aware monitoring beyond generic threshold systems. Network analytics and graph based algorithms uncover hidden relationships, coordinated activities, and transaction patterns that conventional monitoring tools frequently overlook. Real time anomaly detection models process continuous transaction streams using statistical learning and neural network architectures to identify suspicious behavior with greater speed and precision.
The presentation further explores intelligent workflow automation techniques that streamline compliance investigations through automated data collection, alert summarization, and risk prioritization. Ensemble based dynamic risk scoring methods combine transactional, temporal, and contextual attributes to generate more comprehensive threat assessments while improving resource allocation for investigative teams.
In addition to technical implementation strategies, the session addresses practical enterprise challenges including system interoperability, scalable deployment architectures, model governance, and regulatory compliance considerations. Real world examples illustrate how financial institutions can operationalize machine learning driven monitoring systems while maintaining transparency, auditability, and operational sustainability.
Attendees will gain insights into applying AI driven analytics frameworks within modern financial systems, enabling a transition from reactive monitoring toward proactive and intelligent financial crime prevention. The session highlights how scalable data science solutions can strengthen institutional resilience, improve investigative efficiency, and support secure digital financial ecosystems.
Conference Session · Software Engineering and Architecture
Civic Agents in Practice: The Next Phase of AI for Government Workflows
- When
- Friday, October 30 · 1:15 PM–2:15 PM ET
- Room
- Room B
- Speaker
- Himanshu Goil
Public agencies have already experimented with basic AI tools, and many are now asking a more meaningful question: how can AI actually support the work that public staff do every day? This session takes the idea of Civic Agents a step further and shows how they can meaningfully assist staff in their everyday responsibilities, from reviewing requests to organizing information and guiding routine processes.
Instead of focusing on future possibilities, the session highlights what is working today in departments such as finance, permitting, HR, public works, public safety, and administration. Participants will see how Civic Agents can help prepare documents, summarize cases, check information across systems, and reduce the repetitive steps that slow down work.
The focus is practical and grounded. Attendees will learn how Civic Agents fit naturally into existing workflows without requiring major system changes. The session also addresses the key questions public teams often raise, including privacy, access control, and how to introduce AI in a responsible and transparent way.
This session is ideal for government teams that have moved beyond early pilots and want a clear, actionable understanding of the next phase of AI adoption. By the end, participants will be able to recognize strong internal use cases, understand how to prepare their data and processes, and confidently plan the next steps for bringing Civic Agents into their organizations.
Conference Session · Artificial Intelligence and Machine Learning
Designing Conversational BI: How Azure Databricks AI/BI Genie Unlocks Self-Service Insights
- When
- Friday, October 30 · 1:15 PM–2:15 PM ET
- Room
- Room C
- Speaker
- Mou Rakshit
Business users want faster answers from enterprise data, but most self-service analytics still depends on dashboards, SQL requests, or long back-and-forth cycles with data teams. Conversational BI changes that experience, but only if the answers are trusted, governed, and grounded in the right business context.
This session shows how to design a practical conversational BI experience using Azure Databricks AI/BI Genie. We will walk through how to prepare curated datasets, define business-friendly metadata, configure a Genie space, guide users toward reliable questions, and generate accurate answers, summaries, and visualizations from governed enterprise data.
The session will focus on what actually makes conversational BI work in the enterprise: trusted tables, clear metric definitions, helpful synonyms, documented assumptions, row-level security, user feedback loops, and well-designed example questions. Using a demo scenario, attendees will see how non-technical users can ask natural language questions, explore trends, compare performance, and drill into business drivers without writing SQL.
We will also cover common implementation mistakes, including connecting Genie to unclear datasets, skipping semantic preparation, expecting AI to fix poor data quality, and launching without user training or governance review. Attendees will leave with a practical design checklist for building conversational BI that is useful, safe, and adoption-ready.
Conference Session · Artificial Intelligence and Machine Learning
Headroom: A Context Optimization Layer for LLM Applications
- When
- Friday, October 30 · 1:15 PM–2:15 PM ET
- Room
- Room E
- Speaker
- Tejas Chopra
LLM tokens are expensive. With context windows expanding to 200K+ tokens, a single API call can cost several dollars & in production systems handling thousands of requests, these costs compound quickly.
Most optimization efforts focus on model selection or prompt engineering, but the context itself often contains massive redundancy.
Headroom is an open-source Python library (https://github.com/chopratejas/headroom) that sits between your application and your LLM provider, transparently optimizing context before it reaches the model.
The core insight is simple: LLM contexts—especially in agentic workflows—are filled with repetitive tool outputs, verbose JSON arrays, and boilerplate that consumes tokens without adding proportional value
Headroom introduces novel concepts such as reversible compression, cache aligners, compression routers, and even persistent memory
Real-world results:
- 50-90% token reduction on typical agentic workloads
- Drop-in integrations for LangChain, OpenAI, Anthropic, and any OpenAI-compatible provider
- Zero code changes required when using the proxy server
Conference Session · Software Engineering and Architecture
Blending Product Thinking with Software Modernization
- When
- Friday, October 30 · 2:30 PM–3:30 PM ET
- Room
- Room A
- Speaker
- Brian McKeiver
In the world of software, products and systems don’t just remain static, they evolve, accumulate technical debt, and sometimes become an unmanageable mess of outdated frameworks, workarounds, and fragile dependencies.
When dealing with a 15-year-old platform, the question isn’t just how to modernize but what to modernize. Should you refactor piece by piece or completely rebuild? What about adding that new feature for the Product Owner. The answer isn’t always obvious.
This session explores how blending product thinking with software modernization leads to better outcomes. We’ll examine the lens of long-term product strategy, user needs, and business value. Using real-world lessons from a long-running SaaS product, a case study in modernizing a complex legacy platform, we’ll highlight how to balance technical constraints with customer expectations.
Whether you’re leading a legacy transformation or trying to avoid one in the future, this talk will showcase some smart modernization decisions that don’t make the business take a full stop.
Conference Session · Software Engineering and Architecture
Diverge, Converge, Speak Up: The Discipline of Team Creativity
- When
- Friday, October 30 · 2:30 PM–3:30 PM ET
- Room
- Room E
- Speaker
- Tom Rodgers
The best software doesn't come from the best individuals. It comes from teams that know how to think together. Things work when a junior developer can challenge a senior architect's assumption, when a tester can flag that a solution is heading in the wrong direction before it's too late, and when the presence of a VP in the room doesn't freeze the conversation.
That kind of team doesn't happen by accident. It's built through deliberate culture, clear norms, and an understanding of what gets in the way.
This talk explores the conditions that make software teams genuinely creative: what psychological safety really means in practice, how teams move between expanding possibilities and committing to decisions, how to recognise when someone is closing down a conversation that should still be open, and what specific character dynamics, and hierarchy effects, do to team creativity when left unmanaged.
Practical, honest, and grounded in 30 years of building and leading teams.
Conference Session · Software Engineering and Architecture
Don’t Let Your Org Chart Write Bad Code
- When
- Friday, October 30 · 2:30 PM–3:30 PM ET
- Room
- Room C
- Speaker
- Luis Fernandez
In this session, we’ll explore why the biggest driver of your software architecture isn’t technology—it’s the way your teams are structured. We’ll dive into Conway’s Law, which explains how software systems mirror the communication patterns within an organization and look at Inverse Conway Maneuvers to see how reorganizing teams can help align architecture with business goals. To ignore Conway’s law is a trap that will lead to software that is more expensive, and less effective, especially in the micro-services and composable era. We’ll also examine Team Topologies, focusing on how different team types and interaction modes reduce complexity, improve collaboration, and speed up delivery.
Real-world examples will demonstrate how intentional organizational design can eliminate silos and enhance product quality by keeping your codebase closely tied to the customer experience and broader business objectives.
Conference Session · Product, UX, and Digital Experience
Knights, Castles, and Creative Cloud: Rapid Illustration Concepting with AI
- When
- Friday, October 30 · 2:30 PM–3:30 PM ET
- Room
- Room B
- Speaker
- Rex Rainey
Explore how AI can dramatically speed up illustration concepting without replacing the artist’s vision. Using an interior graphics project as a real-world example, we’ll look at how Adobe Firefly enabled rapid iteration of complex visual elements—freeing up more time for refinement, storytelling, and craft.
Conference Session · Cybersecurity, Privacy, and System Resilience
Securing Cyber Physical Systems in Municipal Infrastructure Environments
- When
- Friday, October 30 · 2:30 PM–3:30 PM ET
- Room
- Room D
Cyber physical systems that support municipal infrastructure are becoming increasingly interconnected, creating new opportunities for efficiency while also introducing significant cybersecurity risks. Water treatment facilities, emergency response systems, and public utilities depend on tightly integrated operational technology and information technology environments, making them attractive targets for adversaries seeking to disrupt essential services. This session explores how to secure these complex environments using practical, field tested strategies.
The presentation examines the challenges of protecting systems that must remain continuously available while operating on legacy platforms. It outlines a defense approach centered on network segmentation, where critical control systems are isolated into secure zones to prevent lateral movement during an attack. Attendees will gain insight into implementing monitoring solutions that establish behavioral baselines and detect unusual activity across industrial networks.
Another focus area is the secure integration of remote access capabilities, which are essential for maintenance and operations but often introduce vulnerabilities if not properly controlled. The session explains how strong authentication and controlled access pathways can reduce exposure without limiting operational efficiency.
The discussion also highlights the importance of aligning technical controls with broader resilience strategies, including incident response planning, risk assessments, and adherence to established cybersecurity standards. These elements work together to ensure that systems can withstand and recover from disruptions caused by cyber incidents.
Designed for practitioners working in IT, engineering, and infrastructure domains, this session provides clear, actionable insights that can be applied immediately. Attendees will gain a deeper understanding of how to secure cyber physical systems while maintaining performance, reliability, and safety in mission critical environments.
Conference Session · Artificial Intelligence and Machine Learning
Building Scalable Multi-Agentic AI Systems: Orchestrating Agents with Event-Driven Approach
- When
- Friday, October 30 · 3:45 PM–4:45 PM ET
- Room
- Room E
- Speaker
- Mary Grygleski
This talk will guide developers through the design and implementation of multi-agent generative AI systems using event-driven principles. Attendees will learn how autonomous GenAI agents collaborate, communicate, and adapt in real-time workflows using modern frameworks (such as LangChain or Haystack) and messaging protocols.
Cover core patterns such as pub/sub, orchestrator, and supervisor for managing agent communication, memory, and workflow state. Discuss interoperability with MCP (Model Context Protocol) and how event streaming tools like Kafka, Pulsar, that can facilitate high throughput and reliability.
Conference Session · Artificial Intelligence and Machine Learning
Case Study: Designing AI-Augmented Enterprise Investigation Systems
- When
- Friday, October 30 · 3:45 PM–4:45 PM ET
- Room
- Room A
- Speaker
- Santosh Vasudevan
Organizations today face growing operational complexity, fragmented enterprise data, and increasing pressure to investigate anomalies, incidents, and high-risk events faster than traditional workflows allow.
This interactive case study session explores how modern AI architectures can augment enterprise investigation systems by accelerating evidence gathering, cross-system analysis, policy retrieval, and structured decision support while maintaining human accountability and oversight.
Using a realistic enterprise investigation scenario, attendees will walk through how a traditional investigation workflow operates, where operational bottlenecks emerge, and how AI-assisted systems can improve speed, consistency, and visibility across complex environments.
The session focuses on practical implementation patterns rather than theoretical AI concepts, including:
* AI agents for multi-step investigation workflows
* Retrieval-Augmented Generation (RAG) for organizational knowledge retrieval
* Enterprise tool orchestration and system integration
* Explainable AI outputs and structured reasoning
* Human-in-the-loop decision support models
* Designing AI systems that augment expertise instead of replacing it
Attendees will participate in guided workflow redesign exercises and group discussions to evaluate where AI provides meaningful operational leverage and where human judgment remains essential.
The session is intentionally industry-agnostic and applicable across operations, cybersecurity, risk management, enterprise support functions, compliance, and other investigation-heavy environments.
Attendees will leave with practical frameworks, architectural ideas, and governance considerations they can apply when designing AI-augmented enterprise workflows inside their own organizations.
Conference Session · Artificial Intelligence and Machine Learning
Digital Transformation or Digital Chaos? A Practical Guide for Leaders Who Are Done Playing Catch-Up
- When
- Friday, October 30 · 3:45 PM–4:45 PM ET
- Room
- Room B
- Speaker
- Matt Russell
Most organizations don't fail at digital transformation because of bad technology. They fail because they started with the tool instead of the problem. After nearly 30 years leading technology modernizations — from a 10-person shop to the U.S. Federal Government — Matt Russell has seen the pattern repeat itself at every scale.
This session cuts through the noise with a frank, data-driven look at what actually works: how to audit your digital foundation, make smarter technology decisions without analysis paralysis, manage change when your team isn't buying in, and position your organization for an AI-driven future that's already underway.
You'll leave with a practical decision framework, a clear picture of where AI fits in your operations today, and several concrete actions you can take before the week is over.
No jargon. No vendor pitches. Just hard-won lessons from the field — and a few laughs along the way.
Conference Session · Advanced Manufacturing and Industry 4.0
How to get started with collection and storage of industrial data
- When
- Friday, October 30 · 3:45 PM–4:45 PM ET
- Room
- Room D
- Speaker
- Grant Pinkos
Data collection from industrial devices is often the least understood and most daunting task that a plant can face, yet having that data can give clarity on OEE, maintenance issues, utility usage, and many hard-to-spot trends. Temperature controllers, compressors, variable frequency drives, industrial cooling systems, and sensors that transmit data in any number of protocols (often archaic) will be explained and demystified in this session. The attendees will learn about the tools and data storage systems they can quickly configure at their plant to begin collecting, storing, analyzing and alerting on their data.
Conference Session · DevOps, Cloud, and Platform Engineering
Test Smarter with AI: Playwright Automation at Scale on Azure
- When
- Friday, October 30 · 3:45 PM–4:45 PM ET
- Room
- Room C
- Speaker
- Randy Pagels
Manual testing slows delivery and limits confidence in release quality. In this session, you'll see how AI bridges the gap by turning manual test plans into automated Playwright scripts using GitHub Copilot’s Model Context Protocol (MCP). Combined with Microsoft’s Azure App Testing Service, we’ll show how to run these tests at scale across browsers and platforms. The result is faster feedback, fewer regressions, and more resilient applications. If your team is ready to reduce testing bottlenecks and embrace intelligent automation, this session is your next step.