Moonshot AI MLE Interview Questions (2026)

Questions reported from Moonshot AI machine learning engineer interviews. This bank tracks 11 questions asked at Moonshot AI (8 with free model answers). Practice them live with DeepOffer's AI voice interviewer.

← All 465 questions

Practice this question with an AI interviewer

Get asked follow-ups live, then receive a scored report — like a real MLE interview loop.

Start AI mock interview

#45 Explain multi-head latent attention and the motivation DeepSeek had for introducing it. ★
#49 Explain rotary position embeddings and how position interpolation or YaRN stretches context past the training length. ★
#51 Describe mixture-of-experts and how it grows model capacity without growing per-token compute. ★
#54 Explain the lost-in-the-middle phenomenon in long contexts and how you would mitigate it. ★
#68 Explain what breaks first when pushing Kimi from 8K to hundreds of thousands of tokens, and why.
#69 Explain Kimi K2's MLA and compare it with GQA for KV-cache reduction.
#165 Contrast the prefill and decode phases and explain why prefill is compute-bound while decode is memory-bandwidth-bound. ★
#169 Explain prefix/prompt caching: when to use it and what invalidates a cached prefix. ★
#179 Explain disaggregated prefill/decode serving and when it pays for itself. ★
#232 Explain when a 1M-token context window is the right tool versus retrieval for a long-context assistant.
#299 Hand-write a KV cache and single-step decode. ★