The 60-second answer
Define the task and failure modes before choosing prompting, RAG, fine-tuning, or an agentic architecture. Separate evaluation into task quality, factuality/grounding, safety, latency, and cost rather than one blended score.
Build the answer in this order
1
Define the mechanism
Define the task and failure modes before choosing prompting, RAG, fine-tuning, or an agentic architecture.
2
Explain the architecture
Separate evaluation into task quality, factuality/grounding, safety, latency, and cost rather than one blended score.
3
Compare trade-offs
Use representative held-out prompts, deterministic checks where possible, and calibrated human review for subjective criteria.
4
Close with serving + evaluation
Monitor slice-level regressions and trace retrieval/tool/model behavior after deployment.
A useful interview mental model
This is the shape of a strong answer—not a script to memorize.
01Definition
02Mechanism
03Trade-offs
04Failure modes
05When to use
Senior-level signal
- LLM-as-judge can scale evaluation but requires bias checks, reference examples, and periodic human auditing.
- Plan fallbacks, versioning, and rollback because model/provider changes can shift behavior without code changes.
What the interviewer is really testing
Understanding beyond prompting: architecture, retrieval, evaluation, inference, safety, latency, cost, and failure recovery.
Likely follow-up questions
What assumption makes this approach work?
When would you choose the strongest alternative instead?
What production or data failure mode changes your answer?
Common weak-answer patterns
- Reciting a definition without mechanism or assumptions.
- Claiming one technique is always better without a data regime.
- Stopping before failure modes, validation, or deployment implications.