Talks on shipping AI that actually works in production.
I speak to engineering teams and conferences about agentic systems, RAG architecture, and eval-driven development — grounded in what's actually shipped, not slideware.
Multi-Agent Systems in Production
Orchestrating LangGraph agents on AWS Bedrock — state, reducers, and the failure modes nobody warns you about.
LLM-as-a-Judge: Eval-Driven AI Dev
Building automated quality gates into CI/CD so AI-generated code and outputs get caught before they ship.
RAG That Survives Contact With Users
Lessons from regulated, high-traffic RAG pipelines — chunking, retrieval quality, and hallucination guardrails.
AI-Assisted Engineering at Scale
What changes on a team when Claude Code and Cursor become default tooling — velocity, review, and risk.
Shipping AI in Regulated Industries
HIPAA-aware GenAI features and defense-grade platforms — what compliance actually constrains in practice.
Have something specific in mind?
Happy to tailor a talk to your event's theme or audience level — just ask.