- The session will start with introducing an ideal path to production
--> Choosing LLM (choices, design considerations)
--> Choosing frameworks (choose between Open source frameworks like LangGraph, Autogen, etc.)
--> Choosing architecture (Single/Multi-Agent)
--> Evaluation Strategy
--> Implement Guardrails
--> Scaling Agents
--> AgentOps
- Next, take a deep dive into Evaluation techniques in Agentic AI
--> LLMs as a Judge
--> Human-in-the-loop
- Next, discuss a few Evaluation frameworks
--> Azure AI Evaluation SDK
--> Mosaic AI Evaluation
--> Evaluating with Azure Foundry
- Evaluating Risks with Red Teaming Agents
- Lastly, how do you easily scale out your application with DAPR