AI Learning Jam #2
We're done talking about AI. This is about building it, breaking it, and understanding why it breaks.
AI Learning Jam - In Person #2 is about pushing past tutorials into real context-aware systems and production-grade thinking.
No panels. No talks about whatever's trending in the news cycle this week. Just hands-on work: you'll build a RAG system that understands your actual codebase, then spend the second half building and breaking guardrails on an existing AI application.
What you'll actually do:
Part 1: RAG that knows your codebase inside out
We're vectorizing an entire codebase, not just for search, but so your LLM can pick up on patterns, conventions, and the unwritten rules your team actually follows. You'll use LangChain to build a code review assistant that gives opinions grounded in how your team actually writes code, not generic best practices pulled from the internet.
Part 2: Build guardrails, then try to break them
We switch to a different AI application. Your job: layer in guardrails (system prompt, input validation, output filtering) while others in the room try to red-team it. Prompt injection, edge cases, whatever you can think of. Then you'll see what actually breaks and why hardening matters before you ship.
Who should come:
Anyone curious about building with AI. Beginners welcome. Experts welcome. If you want to learn by watching and doing, you'll fit in.
What to bring:
Your Laptop. That's it. We'll walk you through using free, open-source and lightweight tools for everything else.
Note - To move things faster, you can sign up to OpenRouter.
Schedule:
2:30 PM - Check-in & Networking
2:45 PM - RAG Deep Dive: Vectorizing codebases, OpenRouter, LangChain, production patterns
4:00 PM - Guardrails & Red-Teaming: Build defenses, break them, understand what matters
5:00 PM - Wrap
See you there!