Start here: AI Engineering

English · Русский · AI Interview Atlas › Tracks and roles

The priority questions of this track, in study order. Answer each one aloud before you read anything: the theme page behind every question adds what a strong answer covers and what to read.

For a study sequence with weekly practice and completion checks, use the learning path.

Choose a focused route from the learning roadmap: AI-assisted coding, production AI engineering or AI Leadership.

See the coverage audit, Engineering research, Leadership audit and expansion plan for evidence limits and next steps.

Answers to these questions

• 1. Derive scaled dot-product attention and explain how its scale affects softmax gradients.
  • Knowledge · LLM fundamentals · ✍ Answer
• 2. Derive the memory required by a decoder's KV cache as batch size and context length grow.
  • Knowledge · LLM fundamentals · Asked at: Amazon †, OpenAI † · ✍ Answer
• 3. Compare multi-head, multi-query and grouped-query attention in quality and serving memory.
  • Knowledge · LLM fundamentals · Asked at: Meta † · ✍ Answer
• 4. Under a fixed training compute budget, how do Chinchilla-style scaling results change the balance between model size and training tokens?
  • Knowledge · LLM fundamentals · Asked at: Anthropic † · ✍ Answer
• 5. How would you test and reduce a model's tendency to miss relevant evidence in the middle of a long context?
  • Knowledge · LLM fundamentals · ✍ Answer
• 6. Contrast prefill with autoregressive decode and explain when compute or memory bandwidth becomes the bottleneck.
  • Knowledge · Inference, serving and cost · ✍ Answer
• 7. Explain iteration-level batching and how requests enter and leave a running inference batch.
  • Knowledge · Inference, serving and cost · Asked at: Anthropic † · ✍ Answer
• 8. Explain exact speculative decoding, its acceptance step and workloads where the draft model adds overhead instead of speed.
  • Knowledge · Inference, serving and cost · ✍ Answer
• 9. Estimate a serving memory budget for a 70-billion-parameter model, explicitly choosing precision, context and concurrency assumptions.
  • Applied scenario · Inference, serving and cost · ✍ Answer
• 10. You are asked to reduce serving cost by an order of magnitude. Rank the levers and explain how you would test whether that target is achievable.
  • Applied scenario · Inference, serving and cost · Asked at: Amazon †, Microsoft † · ✍ Answer
• 11. After a deployment, p99 latency doubles although model weights are unchanged. How do you isolate the cause?
  • Applied scenario · Inference, serving and cost · Asked at: Amazon †, Databricks †, OpenAI †, Perplexity † · ✍ Answer
• 12. Choose a chunking approach for technical documentation, including how you preserve meaning across section boundaries.
  • Applied scenario · RAG and retrieval · ✍ Answer
• 13. When would you combine lexical and dense retrieval, and how would you merge their rankings?
  • Knowledge · RAG and retrieval · Asked at: Microsoft †, Perplexity † · ✍ Answer
• 14. How would you separately evaluate document retrieval and answer generation in a RAG application?
  • Applied scenario · RAG and retrieval · ✍ Answer
• 15. Design retrieval that enforces the source system's access rights, including permission changes and shared caches.
  • System design · RAG and retrieval · Asked at: Databricks †, Microsoft †, Palantir † · ✍ Answer
• 16. Compare exact search, HNSW and IVF-PQ for an embedding index under memory, recall and latency constraints.
  • Knowledge · RAG and retrieval · ✍ Answer
• 17. What does interleaving reasoning with tool observations add to a language-model agent compared with reasoning alone?
  • Knowledge · Agents, tools and protocols · ✍ Answer
• 18. Design recovery from tool errors and timeouts, including cases where a timed-out call may already have changed external state.
  • System design · Agents, tools and protocols · Asked at: OpenAI † · ✍ Answer
• 19. How would you select, name and document tools so that a model chooses the right operation and arguments?
  • System design · Agents, tools and protocols · Asked at: Anthropic † · ✍ Answer
• 20. Define completion and stopping conditions for an agent loop so it cannot spend indefinitely on an unfinished task.
  • System design · Agents, tools and protocols · ✍ Answer
• 21. Design human approval for consequential agent actions, including how approval remains bound to the exact action being executed.
  • System design · Agents, tools and protocols · Asked at: OpenAI †, Palantir † · ✍ Answer
• 22. Walk through a preference-based RLHF pipeline and explain the roles of the reward model and reference-policy penalty.
  • Knowledge · Fine-tuning and post-training · ✍ Answer
• 23. Compare DPO with PPO-based RLHF, including the assumptions behind offline preference learning and reasons to collect fresh rollouts.
  • Knowledge · Fine-tuning and post-training · ✍ Answer
• 24. Derive the low-rank weight update used by LoRA and propose an experiment for choosing its rank.
  • Knowledge · Fine-tuning and post-training · ✍ Answer
• 25. How does QLoRA reduce finetuning memory, and which tensors remain trainable or require higher precision?
  • Knowledge · Fine-tuning and post-training · ✍ Answer
• 26. Choose between prompt changes, retrieval and finetuning for a failing AI feature; include data freshness, latency and total cost.
  • Applied scenario · Fine-tuning and post-training · Asked at: Databricks †, Microsoft †, OpenAI † · ✍ Answer
• 27. Design an LLM judge and a calibration process that exposes order, verbosity and self-preference biases.
  • System design · Evaluation and observability · Asked at: Perplexity † · ✍ Answer
• 28. Design a release gate for prompt and model updates, including what happens when aggregate gains hide a critical regression.
  • System design · Evaluation and observability · Asked at: Anthropic † · ✍ Answer
• 29. Benchmark results improve but users report a worse product. What hypotheses would you test first?
  • Applied scenario · Evaluation and observability · ✍ Answer
• 30. Design observability for a production LLM workflow: which spans, versions, costs and feedback should be linked?
  • System design · Evaluation and observability · ✍ Answer
• 31. How should evaluation of a tool-using agent differ from grading one generated response?
  • System design · Evaluation and observability · ✍ Answer
• 32. Compare projection layers, cross-attention and token-based integration for giving a language model access to images.
  • Knowledge · Multimodal and voice · Asked at: Meta † · ✍ Answer
• 33. How would you improve and evaluate speech recognition for mixed-language utterances and different accents?
  • Applied scenario · Multimodal and voice · ✍ Answer
• 34. Design a knowledge assistant over ten million enterprise documents with per-user permissions and a continuously changing corpus.
  • System design · AI system design · Asked at: Amazon †, Databricks †, Microsoft †, OpenAI † · ✍ Answer
• 35. Design an LLM gateway with provider routing, failover, caching, rate limits and enforceable spending budgets.
  • System design · AI system design · Asked at: Palantir †, Perplexity † · ✍ Answer
• 36. Design a customer-support agent that can execute service actions and transfer the case to a human when needed.
  • System design · AI system design · ✍ Answer
• 37. Design a streaming chat service for hundreds of millions of users, including capacity, conversation storage and graceful overload behaviour.
  • System design · AI system design · Asked at: Anthropic †, Google and Google DeepMind †, Meta †, OpenAI † · ✍ Answer
• 38. Implement scaled dot-product attention with a causal mask and tests that would expose attention to future tokens.
  • Coding · Practical coding · Asked at: Amazon †, Anthropic †, Google and Google DeepMind † · ✍ Answer
• 39. Implement a token-bucket limiter and explain what must change when several workers enforce the same limit.
  • Coding · Practical coding · Asked at: Anthropic †, OpenAI † · ✍ Answer
• 40. Write an asynchronous API batch processor with bounded concurrency, retry jitter and isolated per-item failures.
  • Coding · Practical coding · Asked at: Anthropic †, Perplexity † · ✍ Answer
• 41. How does sparse expert routing increase parameter capacity, and what costs remain despite activating only some experts?
  • Knowledge · LLM fundamentals · ✍ Answer
• 42. Explain how latent attention compresses the KV representation and what must be reconstructed at inference.
  • Knowledge · LLM fundamentals · ✍ Answer
• 43. What makes a prompt prefix reusable, and which changes invalidate or reduce the value of a prefix cache?
  • Knowledge · Inference, serving and cost · ✍ Answer
• 44. How does paging the KV cache reduce waste, and what remains for the scheduler to solve?
  • Knowledge · Inference, serving and cost · ✍ Answer
• 45. A service can spend more on one reasoning run or generate several candidates and verify them. How would you allocate compute under a latency deadline?
  • Applied scenario · Inference, serving and cost · 🧪 generated from job-posting themes · ✍ Answer
• 46. Where should a cross-encoder reranker sit in a retrieval pipeline, and when does its quality gain justify latency?
  • Knowledge · RAG and retrieval · Asked at: Microsoft †, Perplexity † · ✍ Answer
• 47. How would you connect each factual claim in an answer to the retrieved passage that supports it?
  • System design · RAG and retrieval · Asked at: Perplexity † · ✍ Answer
• 48. What does MCP standardise between an AI application and an external service that a model's function-call format does not?
  • Knowledge · Agents, tools and protocols · Asked at: Microsoft † · ✍ Answer
• 49. Design memory for an agent that works across many sessions: what is persisted, retrieved, revised and forgotten?
  • System design · Agents, tools and protocols · Asked at: Anthropic † · ✍ Answer
• 50. When does splitting work among agents improve results, and when do coordination costs dominate?
  • Knowledge · Agents, tools and protocols · ✍ Answer
• 51. How would you measure unsupported claims in a deployed RAG application without treating every fluent answer as correct?
  • System design · Evaluation and observability · Asked at: Anthropic †, OpenAI † · ✍ Answer
• 52. A reasoning summary looks sound, but the answer follows a misleading hint. What should an evaluation and monitoring system conclude?
  • Applied scenario · Evaluation and observability · 🧪 generated from job-posting themes · ✍ Answer
• 53. Design a secure developer API for model access, including authentication, quotas, streaming responses and backwards compatibility.
  • System design · AI system design · Asked at: Anthropic † · ✍ Answer
• 54. Build an end-to-end latency budget for a voice agent, from detecting the end of speech to playing its reply.
  • System design · Multimodal and voice · ✍ Answer
• 55. Implement a minimal agent runner with validated tool dispatch, error handling and a hard step limit.
  • Coding · Practical coding · ✍ Answer
• 56. Distinguish time to first token, time per output token, inter-token latency and throughput when comparing serving systems.
  • Knowledge · Inference, serving and cost · Asked at: Microsoft †, Perplexity † · ✍ Answer
• 57. Why can an exact attention kernel run faster without changing the quadratic attention computation?
  • Knowledge · LLM fundamentals · ✍ Answer
• 58. Design natural-language querying over a warehouse with thousands of tables, from schema selection to safe query execution.
  • System design · AI system design · Asked at: Databricks †, Palantir † · ✍ Answer
• 59. A customer reports worse answers after a model upgrade. How do you verify the regression and decide what to restore or change?
  • Applied scenario · Evaluation and observability · Asked at: OpenAI † · ✍ Answer
• 60. How would you make an agent's production actions reversible where possible and auditable where reversal is impossible?
  • System design · Agents, tools and protocols · Asked at: Palantir † · ✍ Answer
• 61. Budget memory for full bf16 finetuning of a 7B model with Adam, then recalculate for LoRA and explain your optimiser assumptions.
  • Applied scenario · Fine-tuning and post-training · ✍ Answer
• 62. Design training for a model that cannot fit on one accelerator; explain how you partition state and keep communication affordable.
  • System design · Fine-tuning and post-training · Asked at: Google and Google DeepMind † · ✍ Answer
• 63. Explain group-relative policy optimisation and the trade-offs of estimating advantages without a learned value model.
  • Knowledge · Fine-tuning and post-training · ✍ Answer
• 64. Policy reward keeps rising while humans prefer its outputs less. How would you diagnose and limit reward hacking?
  • Applied scenario · Fine-tuning and post-training · ✍ Answer
• 65. Explain how Constitutional AI uses principles, self-revision and AI-generated preferences, and identify what human judgement it still depends on.
  • Knowledge · Fine-tuning and post-training · Asked at: Anthropic † · ✍ Answer
• 66. Explain hypothetical-document retrieval and how you would establish whether it improves your query distribution.
  • Knowledge · RAG and retrieval · ✍ Answer
• 67. In an interview where you are expected to use AI tools, how do you stay accountable for code you did not write by hand, and where would you decline what the model produced?
  • Applied scenario · Practical coding · Asked at: Canva ✅ · ✍ Answer
• 68. A team attributes better results to a reasoning model, but it also changed prompts, tools and compute budget. Design a comparison that identifies what improved.
  • Applied scenario · Evaluation and observability · 🧪 generated from job-posting themes · ✍ Answer
• 69. Distinguish outcome rewards, process supervision and inference-time verification. What evidence would justify each for a new reasoning task?
  • Knowledge · Evaluation and observability · 🧪 generated from job-posting themes · ✍ Answer
• 70. A reasoning API request consumes tokens but returns no usable answer. How would you diagnose it and design a bounded recovery policy?
  • Applied scenario · Inference, serving and cost · 🧪 generated from job-posting themes · ✍ Answer
• 71. Plan a product with the interviewer, then build it in about two hours with the AI coding tools of your choice and defend its design, code and path to production.
  • Applied scenario · Practical coding · Asked at: Sierra ✅ · ✍ Answer
• 72. Review and improve a colleague's draft pull request that adds a cross-cutting feature to an unfamiliar codebase, working alongside coding agents.
  • Applied scenario · Practical coding · Asked at: Sierra ✅ · ✍ Answer
• 73. Inside a clone of the product's repository, implement a data structure the product actually uses, asking for AI help only on targeted syntax questions.
  • Coding · Practical coding · Asked at: Cursor (Anysphere) 🗣 · ✍ Answer
• 74. In an AI-assisted coding session, explain which models you use and why, how you watch the token budget, and how you give the agent its context and task.
  • Applied scenario · Practical coding · ✍ Answer
• 75. How do schema-constrained responses differ from function calls, and which component actually executes an action?
  • Knowledge · Agents, tools and protocols · ✍ Answer
• 76. A long-running agent confidently pursues the wrong objective. How do you locate where its task state diverged and recover?
  • Applied scenario · Agents, tools and protocols · ✍ Answer
• 77. For a coding agent, how would you separate the model's contribution from the harness's contribution to reliable task completion?
  • System design · Agents, tools and protocols · Asked at: Anthropic † · ✍ Answer
• 78. How would you build a useful evaluation set when reference answers are missing and expert time is scarce?
  • Applied scenario · Evaluation and observability · ✍ Answer
• 79. How would you change evaluation when an LLM workflow moves from suggesting actions to executing them in production?
  • System design · Evaluation and observability · Asked at: Palantir † · ✍ Answer
• 80. How would you version prompts and dependent configuration so an observed regression can be rolled back reproducibly?
  • System design · Evaluation and observability · ✍ Answer
• 81. Design a pipeline that extracts structured fields from ten million scanned documents and makes uncertain results reviewable.
  • System design · AI system design · Asked at: Palantir † · ✍ Answer
• 82. Design continuous index updates, including changed documents, deletions and queries during a rebuild.
  • System design · RAG and retrieval · Asked at: Perplexity † · ✍ Answer
• 83. An agent's context fills up over a long task. How would you decide what stays in the window, what is summarised and what moves out of it?
  • System design · Agents, tools and protocols · 🧪 generated from job-posting themes · ✍ Answer
• 84. Several product teams share one internal AI gateway. How would you keep one team's traffic, data and caches from affecting another's?
  • System design · AI platform and operating model · 🧪 generated from job-posting themes · ✍ Answer
• 85. A provider announces that a model version your products depend on retires in three months. How would you run the migration across the teams that use it?
  • Applied scenario · AI platform and operating model · 🧪 generated from job-posting themes · ✍ Answer
• 86. Your AI platform's latency and availability depend on an external model provider. What would you promise internal teams, and how would you keep that promise?
  • System design · AI platform and operating model · 🧪 generated from job-posting themes · ✍ Answer
• 87. Before launching an assistant that can use tools, how would you organise adversarial testing, and what would a clean result let you claim?
  • System design · Safety, security and governance · 🧪 generated from job-posting themes · ✍ Answer