Skip to content

The inference engineering curriculum

The question bank answers what to practise. These answer what to learn, and in what order. Each course is a reading path through the work an inference engineer actually does, with the questions and concepts that go with it. Free lessons in every course, no card needed.

Worked answers on inference mechanisms, with calculations, implementation details, and follow-up questions.

START HERE, FREETurn a latency target into a request contract10 min · no account needed→

The main path

Work these in order.
intermediate3 modules · 6 lessons · ~1.5 hours2 free to read

Serving under pressure: from admission to release

Build a defensible serving policy for one specified worker: reserve KV pages, protect token deadlines, reclaim cancelled work safely, and decide whether a change deserves release.

  1. 01Make admission a promise2/2 free
  2. 02Keep execution and ownership consistent2
  3. 03Defend the release2
READ FREETurn a latency target into a request contract10 min →
advanced3 modules · 6 lessons · ~1.5 hours2 free to read

Reading inference engines: prove what the code preserves

Follow a token through an inference engine, test the state that survives each step, and audit the evidence behind a performance claim. Six lessons turn source reading into a reproducible engineering review.

  1. 01Establish the reference2/2 free
  2. 02Preserve state across steps2
  3. 03Audit the evidence2
READ FREEFollow one token through five ownership boundaries10 min →
ALONGSIDE THE COURSES
The must-know questions →The concept curriculum →Interview prep by company →All inference interview questions →