AAI Agentic Handbook
Interview prep

Prepare like a staff engineer.

The content is the same at every level — what changes is the shape of the answer. Learn the delivery, practise the questions volume by volume, keep a bank of stories, and drill the coding round.

01

How a staff engineer sounds

The four-beat answer, say-this-not-this pairs and the scope signals interviewers listen for.

02

STAR story bank

26 worked behavioural stories tagged by the signal each one proves.

03

Lines to have ready

Openers, closers and what to say when you don't know.

04

Coding round

130 Python problems with tested solutions, complexity and a staff follow-up.

Question banks

140 system questions across four volumes

Vol I

How LLMs work

Tokens, transformers, prompting and why models make things up.

10 questions
Vol I

Training, tuning & inference

Pretraining, fine-tuning, RLHF, and the economics of serving a model.

10 questions
Vol II

RAG Fundamentals

Why RAG exists, when not to build it, and what happens inside a query.

10 questions
Vol II

Chunking & Embeddings

The least glamorous section and the one that decides your ceiling.

10 questions
Vol II

Retrieval & Ranking

Dense vs sparse, hybrid, rerankers, top-k and the vocabulary interviewers check for.

10 questions
Vol II

Production RAG Systems

Multi-tenancy, security, caching, monitoring and incidents — judgment under constraint.

10 questions
Vol III

Agent fundamentals

What makes something an agent, tool use, the agent loop, and when a workflow is better.

10 questions
Vol III

Agents in production

Reliability, cost, evaluation, security and human oversight for agents that run for real.

10 questions
Vol IV

Evaluation

Golden sets, rubrics, LLM-as-judge, human review and the statistics behind a real improvement.

10 questions
Vol IV

LLMOps

Gateways, versioning, cost, rollouts, red teaming and model upgrades.

10 questions
Vol V

AI system design

Design ChatGPT-style serving, an eval platform, a coding agent, enterprise search — with numbers.

10 questions
Vol V

ML infrastructure & serving

Parallelism, GPU scheduling, multi-LoRA serving, autoscaling and prefix-aware routing.

10 questions
Vol V

Estimation & debugging drills

Back-of-envelope maths and production debugging scenarios, timed like the real round.

10 questions
Vol V

Behavioural & leadership

Conflict, influence, killing a project, owning a failure — answered at Mid, Senior and Staff.

10 questions