Start with the person, the task, and what cannot go wrong. A useful system begins with a useful question.
Choose your pathFrom the customer’s problem to production
Build AI systems.
Explain them with confidence.
Learn the foundations, work through real designs, and practise the decisions interviewers test.
Interview this week? Go straight to the final-week review.
A little practice, right here01 / 03
- Write
- Compare
- Score
Internal knowledge assistant
Before you pick a model…
A customer wants an assistant for sensitive internal documents. What would you clarify first?
A quick preview. Your draft stays on this page.
- 35
- ready curriculum modules
- 33
- worked system designs
- 22
- technical worksheets
From a question to a system.
Follow one idea from the customer’s desk to the architecture behind the answer.
Two scenes. One connected idea.
Follow the request through identity, retrieval, and an answer with evidence. Explain why each part belongs.
Explore worked designsScroll to follow the idea, or choose a scene.
Understand the problem
Start with the person, the task, and what cannot go wrong. A useful system begins with a useful question.
Choose your pathDesign the system
Follow the request through identity, retrieval, and an answer with evidence. Explain why each part belongs.
Explore worked designsNext: test the system, practise your explanation, and prepare for the interview.
Find Your Next Step
Choose Your Path
Answer a few questions and get a study path that fits where you are today.
Learn
The curriculum: tracks, modules and lessons, with progress saved in this browser.
Design
The casebook: complete system designs to read as interview answers or deep dives.
Practice
Worksheets and behavioural questions you answer, then score against a model answer.
Interview
Technical worksheets, full mock rounds and the final-week review for the days before the loop.
Your reading and practice progress stays in this browser as you go.
Who It Is For
Beginner
Start from the foundations and follow one path, so you always know what comes next.
Experienced software engineer moving into AI
Skip what you know and spend your time on retrieval, agents, evaluation and production AI.
Data engineer
Build on pipelines and governance you already run, and learn where models change the design.
ML engineer going to production
Move from a working model to a system a customer can trust: serving, evaluation, security and operations.
Architect or consultant
Sharpen the judgement calls: which use case to fund, what to build first, and how to scope it.
FDE candidate
Practise the full loop: discovery, design, behavioural stories and a final-week review.
Revision-only candidate
Interview soon? Go straight to condensed reviews, worked designs and timed practice.
What Makes It Different
A Judgement Layer
Every design asks why, not only what: which constraint bites, which trade-off you accept, and what you would tell the customer.
A Casebook at Three Depths
Read a case as the answer you would say out loud, as the technical deep dive behind it, or as a quick review the night before.
Rubric-Scored Practice
Worksheets make you write the answer first, then check it against what a strong answer covers and score yourself honestly.
Built for Spoken Answers
Interviews are spoken. Material leads with the line you could say, then adds depth only where it earns its place.
What Is Ready Today
The curriculum has 9 Learn tracks and the Interview area, 50 modules in all. Not all of them have material yet, and the site says so on every module page rather than hiding the gaps.
- modules ready
- 35
- modules partly ready
- 9
- modules coming soon
- 3
- modules on hold
- 3
Live today: 22 technical worksheets, 7 behavioural worksheets, 33 worked system designs, and 18 final-week reviews.
Know What to Study Next
Pick a path if you want a sequence. Open the curriculum if you would rather choose for yourself.