INDEX · NOTES

Working notes on AI engineering.

What I'm learning about building with language models, kept in the open and updated as I go. The sections run in order: how models work, how you feed them, how agents are built from them, how you run them in production, and where they're applied. For polished essays, see Writing. Every note credits its sources; here's how.

PART 1 · FOUNDATIONS

  1. 01 · 6 OF 6 TOPICS WRITTENLLM fundamentalsHow language models work and what they can and can't do.
  2. 02 · 7 OF 7 TOPICS WRITTENPrompting and context engineeringDeciding what the model sees: instructions, examples, state and memory.
  3. 03 · 8 OF 8 TOPICS WRITTENRAG and retrievalBringing outside knowledge into the context window.

PART 2 · BUILDING AGENTS

  1. 04 · 6 OF 6 TOPICS WRITTENAgentsWhat an agent is: a model choosing its own actions in a loop.
  2. 05 · 6 OF 6 TOPICS WRITTENAgentic design patternsReusable ways to compose LLM calls and agents.
  3. 06 · 9 OF 9 TOPICS WRITTENHarnessesEverything around the model that turns it into a working agent.
  4. 07 · 8 OF 8 TOPICS WRITTENFrameworks and SDKsLibraries for building agents, compared, and when to skip them.

PART 3 · RUNNING IN PRODUCTION

  1. 08 · 8 OF 9 TOPICS WRITTENEvals and observabilityKnowing whether a system works, and why it failed when it doesn't.
  2. 09 · 7 OF 7 TOPICS WRITTENSafety and guardrailsKeeping models and agents within safe limits.
  3. 10 · 8 OF 8 TOPICS WRITTENInference and servingRunning models fast and affordably.
  4. 11 · 6 OF 6 TOPICS WRITTENFine-tuning and adaptationChanging the model itself, and when that beats prompting or RAG.

PART 4 · APPLIED AND REFERENCE

  1. 12 · 6 TOPICS PLANNEDHealthcare AIApplying all of the above where mistakes cost the most.
  2. 13 · 4 TOPICS PLANNEDReading list and glossaryPapers and articles worth reading, plus quick definitions.