Interview: 'you're given 100 hours of expert data + can simulate — how do you train an agent?'
hardAnswer
- (1) BC first: quick baseline from 100 hours of demos → benchmark.
- (2) Warm-start RL from BC weights → faster than from scratch.
- (3) In sim: PPO or SAC with domain randomization; expert data as replay seed.
- (4) Iterative: sim-fine-tune → deploy → collect new demos on failure cases → SFT round.
- (5) HITL: BC + RL fine-tune with expert intervention on failure states (DAgger-style).
- (6) Monitor gap between BC + RL policies; if RL not exceeding, reward is misspecified.
Check yourself — multiple choice
- Just BC
- BC baseline + warm-start RL from BC + PPO/SAC in sim with DR + expert as replay seed + iterative DAgger with HITL + monitor RL vs BC gap for reward specification
- Random
- Only sim
Expert demos + sim: BC warm-start + RL fine-tune + DAgger + iterative refinement.
#interview
Practise Reinforcement Learning
214 interview questions in this topic.
Related questions
- Interview: 'design an RL system for X' — what's your framework?
- Interview: 'when should you NOT use RL?'
- Interview: 'your DQN agent isn't learning — how do you debug?'
- Interview: 'explain why PPO is on-policy but uses a ratio to importance-sample.'
- Interview: 'walk me through the tradeoffs of SFT-only vs SFT+RLHF vs SFT+DPO.'
- Interview: 'how do you make an RL policy safe to deploy?'