Category · 8 articles

ML & AI Engineering Interviews

The interview lane nobody else covers.

Retrieval pipelines, eval harnesses, model-serving design, LLM system tradeoffs, vector index choice, and cost-shaped scaling for AI engineers.

  • ml engineer interview
  • ai engineer interview
  • llm system design
  • rag interview
  • ml system design
  • vector database interview
18 min read3,631 words

Fine-Tuning vs Prompting: Economic and Operational Tradeoffs in Interviews

Pick a lane based on data volume, latency, and maintenance — not hype.

  • fine tuning interview
  • llm finetuning interview
Read article
18 min read3,596 words

Feature Stores and Training-Serving Skew: What Staff Interviews Probe

Schema drift silently kills models — show you instrument for it.

  • feature store interview
  • ml platform interview
Read article
18 min read3,637 words

RAG System Design Interviews: Chunking, Embeddings, and Grounding

Retrieval quality is the product — generation is only half the story.

  • rag interview
  • llm system design
  • ml system design
Read article
18 min read3,630 words

LLM Serving: Latency, Batching, and Cost per Token in Interviews

Interviewers want production sense — not a diagram of every GPU feature.

  • llm serving interview
  • ai engineer interview
Read article
18 min read3,634 words

Safety Layers for LLM Apps: PII, Prompt Injection, and Policy

You will not solve abuse with a regex — but you can show mature thinking.

  • prompt injection interview
  • llm safety interview
Read article
18 min read3,622 words

Vector DB Interviews: HNSW vs IVF, Filters, and Operational Reality

Pick assumptions — vendor marketing is not a substitute for workload fit.

  • vector database interview
  • embedding search interview
Read article
18 min read3,612 words

ML Evaluation Interviews: Metrics, Leakage, and Online-Offline Gaps

Offline accuracy without methodology is a vanity metric — say why your split is valid.

  • ml engineer interview
  • model evaluation interview
Read article
11 min read1,534 words

ML System Design Interview: A RAG Pipeline, End to End, in 45 Minutes

How to design a retrieval-augmented generation system at L5/L6 depth without losing the room.

  • ml system design
  • rag interview
  • llm system design
  • ml engineer interview
Read article

Stop grinding. Start patterning.

Patterns-first interview prep — coding, design, behavioral, mocks, ML/AI in one $19/mo subscription.

See the full loop →