AI interview & certification prep questions, every answer referenced
1100+ questions built from real AI talks and lectures — by working engineers, researchers and university faculty — each one written from what the speaker actually teaches, not a recycled question bank. Follow any answer back to the exact moment it came from and check it yourself.
Transformers, RAG, agents, prompting, evals, fine-tuning and system design — at three levels, so you can start where you actually are.
Already purchased? Restore your access
What a question looks like
A leading open AI model was trained on roughly 10 terabytes of internet text. How big is the finished model?
- AAbout 140 gigabytes — roughly 100 times smaller than the training text
- BAbout 10 terabytes — the same size as the text it was trained on
- CAbout 5 gigabytes — small enough to fit on an ordinary USB stick
- DAbout 1 terabyte — a little smaller than the original training text
From Andrej Karpathy · [1hr Talk] Intro to Large Language Models
Written from 110 talks and lectures across 21 sources — AI researchers, university faculty and working engineers. Every question links to the moment it came from.
AI & LLM Essentials
800+ questions, unlocked once for $4.99 — no subscription, and free questions in every pack.
Compiled from real ai talks and lectures by working engineers, researchers and university faculty. Not affiliated with or endorsed by any university, channel, creator or certification program. Claude and Claude Certified Architect are trademarks of Anthropic; we reference them only to say what this practice material is for.
Claude Certified ArchitectPreparing for the certification
Scenario practice for the Claude Certified Architect – Foundations exam, arranged by its five domains.
Agentic Architecture & Orchestration · Claude Code Configuration & Workflows · Tool Design & MCP Integration · Context Management & Reliability · Prompt Engineering & Structured Output · Prompt Engineering · AI Agents & Tool Use
7 topics · 323 questions · 36 free to try
FoundationsStart here if AI is new to you
How the models actually work — the mechanics behind every other answer.
LLM Fundamentals · Transformers & Attention · Diffusion & Image Generation
3 topics · 222 questions · 23 free to try · 8-question free warm-up
Building with AIThe AI-engineer core
Retrieval, agents and prompting — what an AI engineer is hired to do.
Retrieval-Augmented Generation · AI Agents & Tool Use · Prompt Engineering
3 topics · 200 questions · 20 free to try · 8-question free warm-up
Shipping & ScalingWhat senior loops dig into
Evals, fine-tuning, inference cost and system design — getting it into production.
LLM Evals & Observability · Fine-tuning & Model Adaptation · Inference & Optimization · GenAI System Design · LLM & Agent Security
5 topics · 263 questions · 26 free to try · 8-question free warm-up
Python Programming
200+ questions, unlocked once for $4.99 — no subscription, and free questions in every pack.
Compiled from university lectures and named practitioners who explain the mechanism, not the syntax. Not affiliated with or endorsed by any university, channel, creator or certification program.
Coming next
Each category is a separate one-time purchase. Votes decide the order.
AI Governance & Security
Alignment, red-teaming, prompt injection and the rules now shaping how AI ships.
Compiled from Policy talks, safety research and practitioner walkthroughs
Classical ML & Statistics
Bias-variance, regularisation, trees and the statistics still asked in every ML and data-science loop.
Compiled from University statistics courses and explainer channels
Would you want these too?
We're focused on AI for now. These aren't planned — but if enough people want one, we'll reconsider.
Algorithms & Data Structures
Complexity, sorting, graphs, dynamic programming.
Databases & SQL
Query planning, indexing, transactions, and the SQL that gets asked.