Notes from the workshop
Practical writing on AI, LLMs, engineering, and running software in production — no hype, no filler.
Evaluating LLM Outputs: Building Your First Eval Set in a Day
A practical guide to building your first LLM eval set in one day: collecting real examples, choosing graders, scoring outputs, and wiring checks into CI.
Vector Databases Compared: When Postgres + pgvector Is All You Need
A practical guide to choosing a vector store — when Postgres with pgvector is enough, when a dedicated vector database earns its keep, and how to decide.
How Much Does It Cost to Build an LLM-Powered Product? An Honest Breakdown
A plain-language breakdown of what it actually costs to build an LLM product — the API bill, the engineering time, and the ongoing costs that never show up in the demo.
Fine-Tuning vs Prompt Engineering: A Decision Guide With Real Trade-offs
A practical guide to choosing between fine-tuning and prompt engineering for an LLM feature — what each actually changes, when to use which, and the costs nobody mentions.
What an AI Agent Actually Is — and When You Don't Need One
A clear guide to what an AI agent actually is, how it differs from a plain LLM call or a fixed workflow, and when a simpler approach is the better choice.
RAG Explained: How 'Chat With Your Documents' Actually Works
A practical, no-hype guide to Retrieval-Augmented Generation — what it is, when it beats fine-tuning, and the parts that actually determine quality.
Choosing a Speech-to-Text Model in 2026: Whisper, Deepgram, and When to Self-Host
A practical guide to picking a speech-to-text model — the trade-offs between open models like Whisper and hosted APIs like Deepgram, and when self-hosting pays off.