Case Studies
Nearly every source agrees this is the round that decides most offers. It goes by a different name per role — case study, system design, incident response, product sense — but the shape is the same: a vague problem, live decomposition, an interviewer pushing back. Pick your role below.
Forward-Deployed Engineer (FDE)
A hypothetical customer hands you a vague problem and you decompose it live, 45–60 minutes — lowest pass rate, highest weight in the loop.
Applied / Agentic AI Engineer
Usually called the system design round: "build an agent that does X," decomposed live under stakeholder pushback.
AI Solutions Architect / Presales Engineer
Often the most heavily weighted round in the loop: three full discovery-to-design walkthroughs across three industries.
SRE / Reliability Engineer for AI Agent Applications
Google calls it NALSD; most other companies call it incident response — triage and design live against a failing system.
AI Product Manager
The AI product sense round: an ambiguous prompt, scoped around a model’s unreliability instead of assumed perfection.
Comments
Comments are provided by Giscus and GitHub. Loading them connects your browser to GitHub, and GitHub's privacy terms apply.