This session walks attendees through a practical progression of building real-world AI systems:

LLM app foundations (prompting, tool use, reliability basics)
RAG done right (retrieval design, chunking, grounding, citations)
Copilots & dev assistants (how to design assistive workflows and context)
Agents & multi-agent systems (planning, tools, memory, orchestration)
Evaluation & guardrails (hallucination mitigation, safety boundaries, monitoring)

Attendees leave with repeatable patterns and a production checklist they can apply immediately—especially in Azure-centric architectures (Azure OpenAI, Azure AI Search, Functions, App Service/Container Apps, Prompt Flow/eval pipelines, observability).

Audience level: Intermediate
Pre-reqs (Required):
- Comfortable with APIs, JSON, and basic backend concepts (auth, logging, retries).
- Familiarity with LLM basics (tokens, temperature, prompt/response).
- Built at least one small LLM prototype or used an LLM SDK.

Pre-reqs (Helpful):
- Basic understanding of vector search / embeddings.
- Familiarity with Azure (App Service/Functions, Azure OpenAI, Azure AI Search) is helpful but not mandatory.