Every article is filed under one of 178 categories. 205 pieces so far.
What the AI engineering role actually is, how it differs from ML engineering, and the vocabulary the rest of the roadmap assumes.
The mechanics under the API: tokens, context windows, sampling, and the knobs that actually change output.
Getting reliable behaviour out of a model with nothing but the message you send it.
Choosing between frontier APIs, open weights, and self-hosting — and the trade-offs behind each.
The wire protocols and client libraries you'll actually write code against.
Turning text, images, and audio into vectors — the substrate for search, RAG, and classification.
Where embeddings live, how they're indexed, and how similarity search behaves at scale.
Grounding a model in your own data: the pipeline, the failure modes, and the frameworks.
Models that plan, call tools, and act in a loop — plus the SDKs that make it tractable.
The open protocol for connecting models to tools, data, and prompts across vendors.
Deciding what goes into the window, what gets compacted, and what never should have been there.
Models that read images, watch video, hear audio, and generate media back.
Knowing whether a change helped — the only thing separating engineering from vibes.
Tracing, cost, and quality signals from production — where the real eval set comes from.
The failure modes that make headlines, and the controls that prevent them.
The editors, CLIs, and coding agents that write the code you review.