Had several rounds with people from various functions, and overall the conversations were thoughtful and engaging. Most of the earlier interviews focused on conceptual thinking, modeling approaches, and communication/problem-solving style. The final stage involved a take-home exercise and presentation, which was worthwhile and relevant to the role. During the presentation, interviewers added comments and questions directly into the document in real time, which made parts of the discussion slightly more difficult to navigate. There was also a strong emphasis on live Python coding and algorithm-style questions, so candidates should be prepared to demonstrate coding fluency under time pressure in addition to conceptual data science knowledge. Overall, the process was organized and communication was responsive throughout. However, I found parts of a post-interview feedback conversation relatively brief and quite direct in tone, and it did not provide much specificity around the evaluation or how the decision was made. While I appreciated the opportunity to discuss it, I would have valued more concrete feedback on areas of mismatch between the role requirements and my performance.

Process (5 rounds): Hiring Manager screen → Take-home project → Take-home extension → CTO interview → CEO interview The Good: The hiring manager was technically sharp, deeply familiar with the Android stack, and genuinely pleasant to talk with. The take-home project was unusually well-designed — working directly with production code — which showed respect for the candidate's time and gave a realistic preview of the work. Strong alignment on technical direction and architecture decisions. The Concerns: The CTO interview was a significant departure in tone. Several questions felt more like challenges than genuine inquiry, and some came across as unnecessarily aggressive for a professional interview setting. More substantively, the engineering culture raised flags. The CTO described a workflow of 20–30 PRs per day across multiple codebases using AI tooling. When asked about code quality practices, the response framed human review as the bottleneck rather than a feature. For a 13-person team owning a large, cross-platform product, that approach to velocity over quality gave me pause — particularly for a sole-platform engineer role where you'd be responsible for long-term maintainability. Takeaway: Great product, strong HM, and a modern tech stack. But if engineering culture and code quality practices matter to you, come prepared to ask pointed questions about how quality is maintained at speed. The answer you get will tell you a lot. At the end, I opted to not continue to the CEO round.