Practice
Select your target and start a session
Target Company
Read the full Anthropic interview guideInterview Type
How MLE / Research interviews work at Anthropic
Unlike SWE phone screens (which often pack multiple sub-problems), MLE rounds typically give you one substantial task and explore it for the full hour. Loops are usually 2–5 rounds.
Common round formats: Debugging (find bugs in NumPy/RL code), Fundamentals (implement attention/NN from scratch), Implementation (build data batcher/sampler), Applied LLM (design classifier with prompt engineering), Perf Modeling (math on paper, no coding), OA Take-home (4-level escalating spec), Research (open-ended take-home, 4+ hours).
Questions
1 shown · Anthropic MLE / Research
Free preview · Try Web Crawler (Coding) and Prompt Playground (System Design)Upgrade →