From “what's a token?”
to staff AI engineer.
The free, level-by-level path through AI engineering — with the interview built in. Every page is written for a level, carries day-one defaults, real numbers and primary sources, and every piece of jargon links to a page with a picture: hovertap a highlighted term to try it.
Reviewed October 2026 · vendor-neutral · sources on every page · how it's written
Start where you are
Five ways in, one handbook
No hands-on yet
You've used ChatGPT or Claude and want to understand what's going on underneath. No code required.
Open the Curious track →BuddingStarting out
Students, new grads and engineers moving into AI. You can code; you want the mental models and your first working build.
Open the Budding track →MidBuilding real systems
You've shipped something with an LLM. Now: retrieval quality, evals, agents that don't loop forever.
Open the Mid track →SeniorRunning it in production
Migrations, cost, latency, incidents, rollouts. The operational layer most tutorials skip.
Open the Senior track →StaffOwning the trade-offs
Multi-tenancy, security, org-wide platforms — and how to say it clearly in a design review or interview.
Open the Staff track →Five volumes
Read in order, or jump to what you need
LLM Foundations
What a language model is, how it is trained, and what actually happens when it answers.
20 questions · 7 postsVol IIRetrieval-Augmented Generation
Giving a model the right knowledge at answer time — chunking, retrieval, ranking, production.
40 questions · 7 postsVol IIIAgents
Models that take actions: tools, loops, memory, multi-agent systems and keeping them safe.
20 questions · 7 postsVol IVEvals & LLMOps
Knowing whether it works, and shipping changes without breaking it.
20 questions · 7 postsVol VSystem Design & Interview Loops
End-to-end AI system design with capacity maths, ML infrastructure, estimation and debugging drills, and behavioural rounds by level.
40 questionsThe RAG pipeline, clickable
Every box is a term page
From the blog
Breaking into AI engineering: a roadmap from zero
The job explained through one wrong answer, then a level-by-level ladder with honest time estimates, a learning order and five things to read.
What happens when you ask an AI a question
From pressing enter to the last word: tokens, the desk the model reads from, one-token-at-a-time writing, thinking, web search, and where your words go.
The three dials: tokens, temperature and context
Two bugs show up in nearly every first LLM integration. Both trace back to the same three settings, and both are fixable in a few lines of code.
Prompt, retrieve or fine-tune? A decision guide
Diagnose the failure first, then pick the technique. A worked claims-triage case shows where fine-tuning pays in 2026: cost, not knowledge.
The real cost of an LLM feature: latency, tokens and the bill
Five terms make up the bill and two make up the wait. Reasoning tokens, cache writes, batch discounts and the self-host break-even, worked with stated assumptions.
What AI-engineering interviewers test at each level
One design question, three scored answers (Mid, Senior, Staff), the rubric behind the scores, and the red flags that sink otherwise good candidates.
Getting ready for an interview?
Staff-level delivery, practised
How a staff engineer sounds
The four-beat answer, phrases to use and avoid.
02STAR story bank
26 worked stories tagged by the signal they prove.
03Coding round
130 Python problems with tested solutions and staff follow-ups.
047-day plan
Day-by-day prep for an AI engineering loop, plus the night-before checklist.