A 1-hour hiring manager (engineering lead) conversation centered entirely on one or more of the candidate's past technical projects. The HM expects the candidate to explain the project in depth: the problem it solved, the technical decisions made, why those decisions were made over alternatives, how they were implemented, and the tradeoffs involved. The HM is particularly interested in projects relevant to the team's domain (e.g., AI agents at Harvey AI). Background fit is heavily weighted — the HM reportedly only engages deeply when the candidate's project aligns with the team's work.
Common mistakes: Choosing a project that is not relevant to the team's domain — HM reportedly loses discussion interest quickly; Inability to explain why specific technical decisions were made; Giving surface-level answers without depth on tradeoffs; Not preparing for a full hour of deep probing on a single project
Interviewer hints: Prep call may signal that the interviewer wants to hear about more cutting-edge methods — take that signal seriously in project selection
What passers do: Selecting a highly domain-relevant project (e.g., AI agent project for Harvey AI HM who works on code-gen agents); Proactively covering why/how/tradeoff framing without waiting to be asked; Demonstrating genuine ownership and depth of understanding; Keeping the conversation engaging and technically substantive throughout the full hour
Why people fail: Being rejected at this round due to poor background/project fit — HM is selective about fit even if other rounds pass; Presenting a project that sounds relevant on the surface but lacks depth when probed; Using an approach that doesn't match what the interviewer was hoping to hear (e.g., not cutting-edge enough for an AI-focused HM)
Edge cases probed: Whether the candidate's project is genuinely relevant to the team's domain; Depth of understanding vs. surface-level familiarity with one's own project; Ability to articulate rejected alternatives, not just the chosen solution
Alternative approaches: Broader multi-project survey (Covering multiple projects shallowly is risky — HM prefers depth over breadth and will lose interest if the project isn't a strong fit. Better to pick one highly relevant project and go deep.)