29 real interview questions · Applied AI / Forward Deployed Engineer
Question bank
0 marked confidentFundamentalsExplain how a large language model works to a smart non-expert.▸
FundamentalsWhen would you use RAG vs fine-tuning vs a longer context window?▸
RAGWalk me through designing a RAG system. Where do most of them fail?▸
RAGHow do you choose a chunking strategy?▸
EvalsHow do you know your LLM feature is actually good? How would you evaluate it?▸
EvalsWhat is LLM-as-judge and what are its pitfalls?▸
AgentsWhen should you build an agent vs a simple prompt chain / workflow?▸
AgentsExplain function calling / tool use and its security implications.▸
ProductionThis feature is too slow and too expensive. How do you fix it?▸
ProductionWhat is prompt injection and how do you defend against it?▸
ProductionHow do you handle hallucinations in a production system?▸
System DesignDesign an AI assistant that answers employee questions over internal company docs.▸
System DesignDesign a system to extract structured data from 100k messy PDFs.▸
CodingIn a coding round you're told: 'call an LLM to classify these 1,000 tickets into categories and output JSON.' How do you approach it?▸
BehavioralTell me about a time you shipped something under significant ambiguity.▸
BehavioralHow do you explain a hard technical tradeoff to a non-technical stakeholder?▸
FundamentalsWhat are the main limitations of LLMs you design around?▸
FundamentalsWhy can't you just trust the model's output directly in a pipeline?▸
FundamentalsExplain the tradeoffs between temperature, top_p, and max_tokens for a customer-facing assistant vs a code-generation tool. (OpenAI signature)▸
Fine-tuningGive me your decision framework for prompt vs RAG vs fine-tune vs distill. (asked everywhere)▸
System DesignDesign a private, VPC-deployed RAG system for a healthcare customer: 50M documents, HIPAA constraints. (Anthropic-style)▸
Customer-FacingYou deployed the platform but adoption is 12% after 90 days, and the customer blames the product. What do you do? (FDE scenario)▸
BehavioralWhy did you use an LLM for that project, and how did you know it was the right fit? (Anthropic HM)▸
FundamentalsWhat is the KV cache and why does it make LLM inference memory-bound?▸
EvalsDesign the eval harness for an AI agent we're deploying to a customer. (2026 design-round staple)▸
ProductionA customer reports the model is 'getting worse' in production. Investigate.▸
Customer-FacingOpen a discovery call with a VP who says 'we want AI in our support flow.' What do you ask?▸
RAGWould you use a vector database for this corpus? How do you decide?▸
CodingHow do you use AI coding tools day-to-day — and how would you use them in this interview?▸