Application — Resume + statement of exceptional work.
Initial Screen/Phone — 15–30 min call (often technical staff/engineer). Focus: background, motivation past projects/challenges, light tech probes. Some report cold/pretentious tone or hard early questions.
Coding/Assessment — Frequently CodeSignal (1 hr, medium-hard LeetCode-style; e.g., array/shape fitting, function implementation). Some roles skip or use live coding.
Technical Rounds/Loop — 2–5 rounds (often virtual "onsite" or spread out):
Coding: Live problems (e.g., filesystem design progressive, async in Python, closures/promises in JS).
System Design: Practical/scaled (e.g., filesystem from scratch, Kubernetes interactions, distributed queues).
ML/AI-Specific (for relevant roles): Project deep-dives, technical challenges, RLHF/dataset creation, LLM experience.
Behavioral: Past achievements, job experience description.
Some mention 4 on-site rounds after screen (e.g., ML Engineer )
Offer — Fast if passed (days/weeks); relocation often supported. Low pass rate implied by selectivity.
