SteveITpro - Learning AI & Cloud in Public

Getting ready for Microsoft Ignite 2025 (Nov 18-21) in San Francisco. This year's focus is heavily on multi-agent AI systems, Azure AI Foundry, and enterprise AI governance - all areas I'm deeply interested in exploring.

Why Ignite 2025?

The enterprise AI landscape is evolving rapidly. Multi-agent systems are moving from research papers to production deployments, and organizations need practical guidance on:

Orchestration - How do you coordinate multiple AI agents working together?
Governance - How do you monitor, secure, and audit AI agent behavior at scale?
Infrastructure - What does the cloud architecture look like for multi-agent systems?

Microsoft Ignite is where theory meets practice - real product teams sharing what's actually shipping, not just research concepts.

📅 Download My Schedule

Download Calendar (ICS) - Import all my sessions into your calendar app (Outlook, Google Calendar, Apple Calendar, etc.)

My Session Plan

I've carefully selected sessions across four major themes:

🤖 Multi-Agent AI Systems

BRK197: Build multi-agent systems with Azure AI Foundry

Microsoft Agent Framework (Semantic Kernel + AutoGen)
Shared state management and Human-in-the-loop patterns
OpenTel observability, MCP toolchains, Activity Protocol
Deploying agents as containers on Azure AI Foundry

LAB513 Series: Agent-framework hands-on

Unified platform from Semantic Kernel + AutoGen teams
A2A (Agent-to-Agent) compatible agents
SWE Agents with GitHub Copilot
MCP tools for enterprise workflows

LAB518: Multi-agent application with MCP and Cosmos DB

Model Context Protocol (MCP) integration
Azure Cosmos DB for agent state and memory
Semantic search for multi-agent systems

Enterprise AI Governance & Security

THR721: AI agent governance at scale

Registering and monitoring all enterprise agents
Preventing agent sprawl and security risks
Centralized governance patterns

LAB516 Series: Automated AI red teaming

Azure AI Foundry safety testing
Automated attack techniques for vulnerabilities
Multi-dimensional risk assessment

THRSP927: Multi-agent compliance automation

SymphonyAI's "Always On Compliance"
Risk-based decisioning with agent ecosystems
Real-world financial services use case

Azure AI Platform & Infrastructure

THRSP904: Azure Databricks + Agent Bricks

Serverless workspace deployment
Lakebase for Data & AI apps
Agent Bricks for no-code data analysis
AI/BI Genie + Copilot Studio in Teams

THR702: Azure SRE Agent

AI-driven cloud operations
Automated detection, diagnosis, remediation
Reducing MTTR with telemetry-to-action AI

LAB517: AKS automation with AI-driven ops

AI alerts for bottleneck detection
Self-healing agents for node management
aks-mcp server for natural language cluster control

Physical AI & Emerging Applications

BRKSP489: Physical AI with NVIDIA + Azure

Simulation-driven robotics development
NVIDIA Omniverse + Azure integration
Digital twins and synthetic data generation

What I'm Looking For

Beyond just attending sessions, I'm interested in:

Patterns over features - What architectural patterns are emerging for multi-agent systems?
Production stories - What works in real deployments vs. what sounds good in slides?
Integration points - How do these technologies actually fit together?
Failure modes - What breaks? What's hard? What are teams struggling with?

Writing Approach

I'll be documenting my learning journey in real-time through this blog:

During the conference: Quick notes and initial reactions
Post-session: Synthesis posts connecting concepts across sessions
Post-conference: Deep dives on specific topics and hands-on experiments

The goal isn't to summarize sessions (Microsoft will have recordings). The goal is synthesis - connecting dots, exploring implications, and sharing what I'm actually going to try implementing.

Following Along

I'll be updating this blog throughout the conference. If you're also attending Ignite or following along remotely, I'd love to hear your perspectives. Find me on LinkedIn.

Let's see what we learn.