Behavioral and Project Deep-Dive (Recruiter Call + HM Round)

Two rounds, three reports. The recruiter call that opens the loop, reported twice in September 2025, and the hiring-manager round of a staff SWE onsite, reported once in October 2025. Both recruiter-call posters described an ordinary, friendly call with no tricky follow-ups and were rejected the next day. The recruiter call is a filter, not a formality.

Given — the questions as reported

Recruiter call, in the order one poster gave them:

• Why Perplexity? Asked first, before anything else.
• What is your most recent proudest project?
• What was your most challenging project, and what made it challenging?
• What is your tech stack?
• Walk me through your background.
• What compensation are you expecting?

Hiring-manager round, staff SWE, described by the poster as standard behavioral:

• How did you resolve cross-team collaboration challenges?
• Give an example of making an unpopular decision.
• How did you handle requirement changes?

Part 1 — the recruiter call

The recruiter also describes the team's daily work and which teams it collaborates with, then the process: a coding interview of 45 minutes where Python is advised, then an onsite of several technical rounds plus a meeting with the manager or CTO. One recruiter said the interview is hard and that the work afterwards has a lot of freedom.

Part 2 — the hiring-manager round

The poster's own outline for each question, prepared with STAR:

• Collaboration: one cross-team project, the communication barrier at the start, your role and responsibilities, and the concrete actions you took, such as setting up alignment meetings and defining ownership.
• Unpopular decision: why it was unpopular at the time, how you communicated and held to it, and how it was later shown to be right.
• Requirement changes: one specific change, how you assessed the impact fast, re-planned, communicated with stakeholders, and still delivered.

The reported wall — the next-day rejection. Neither recruiter-call poster could name a mistake. Commenters offered three readings: many recruiters this year are gathering information and need a hiring manager's go-ahead before any technical round; some calls are scheduled with the rejection already decided; and one poster suspected location, asking whether the other was outside the Bay Area or New York. Treat the call as a real screen and have crisp answers ready for the first two questions.

Context for "why Perplexity," from one commenter. Perplexity differs from the frontier labs: a heavy focus on real-time web indexing and query interpretation, integrating existing models from OpenAI, Anthropic, and Mistral with retrieval rather than training giant models, and a smaller infrastructure footprint with strong emphasis on fast retrieval and grounding. The same commenter called pay on the SWE track generous, which is relevant when the compensation question comes.