Went through the Cohere PM interview loop for a senior PM role last quarter. Background: I came from consulting with B2B SaaS PM experience, no ML background. Sharing specifics on the question types and what seemed to land.
The loop structure for PM
Different from the engineering loop. My version was: 30 min recruiter screen 45 min PM phone screen with a hiring manager (product sense + background) Onsite: product sense round, strategy/market round, cross-functional collaboration round, and a final executive conversation
No take-home for PM. Total loop was about 4 weeks start to finish.
Actual questions I got (paraphrased):
Product sense: How would you think about designing an enterprise dashboard for customers using Cohere's API in production? What are the most important metrics to surface and why? Cohere's customers are mostly developers and ML teams. How does that change how you approach product discovery?
Strategy: How would you assess whether Cohere should build a vertical-specific model for, say, legal or healthcare, vs. keeping the model general and letting customers fine-tune? Who are the top 3 competitors in the enterprise LLM API space and how is Cohere differentiated? (they really tested this)
Behavioral/cross-functional: Tell me about a time you had a major disagreement with engineering on scope or timeline. How did you resolve it? How do you work with researchers who don't necessarily think in terms of product cycles?
What I noticed about fit
They weren't looking for an AI researcher in PM clothing. But they were looking for someone who has genuine curiosity about how the models work and isn't intimidated by technical conversations. You need to hold your own in a room with ML engineers without trying to fake expertise you don't have.