Databricks · Primly Community

Databricks product manager interview questions, full loop breakdown

growth_gabe · 4 replies

Went through the Databricks PM loop in Q1 2026 for a senior PM role on one of their developer tools products. Sharing the questions and format because the PM loop here is different from what you'd expect at a typical SaaS company.

First thing to know: Databricks PMs are expected to be quite technical. Not coding-capable necessarily, but deeply fluent in data infrastructure concepts. If you can't hold a conversation about streaming vs batch processing, or why Delta Lake matters vs a plain data lake, you'll struggle in the product rounds.

Here's roughly what I got:

Intro screen with recruiter. Standard background, why Databricks, logistics. 30 minutes.

Product sense round. "How would you improve the Databricks notebook experience for data scientists who are transitioning from Jupyter?" They want specificity. I laid out user segments (transitioning DS vs. experienced Databricks users), then got into specific friction points. They pushed back on my prioritization, which I think is the point.

Technical depth round. An engineer walked me through a simplified version of a data ingestion problem and asked me to identify the product/design constraints I'd need to surface before committing to a roadmap item. This wasn't a coding test but I needed to understand what tradeoffs the eng team faces.

Execution and cross-functional round. "Tell me about a time you had to ship a product under significant constraint." Very behavioral, but with PM flavor. They probed: how did you scope? What got cut? How did you communicate to stakeholders?

Metrics and outcome round. "How would you measure the success of a new feature for large-enterprise Databricks customers?" I went with retention-based metrics (time-to-value, 30-day activation) plus revenue impact proxies. They liked the layering.

No PM case study / take-home. All live.

My overall impression: they want product people who respect engineering complexity and can translate it for business stakeholders. Pure growth-PM or consumer-PM profiles probably don't fit well without some data infrastructure background.

4 replies

intl_isla

This is really helpful. How technical is "really technical" for the PM role? I have a CS degree but pivoted to PM 4 years ago and my day-to-day is fairly light on infrastructure. Would that be a red flag?

growth_gabe

Honestly it depends on the sub-team. My sense is that if you genuinely understand distributed systems concepts at a high level and can discuss data pipeline trade-offs intelligently, you're fine. You don't need to have written Spark code. But if those concepts feel foreign, that's probably the gap to close before applying.

pm_priya

The notebook UX question is interesting. That's a very real product problem -- Jupyter has 20 years of muscle memory baked into data scientists. I'm curious how deep they let you go on actual UX specifics vs. keeping it high-level strategy.

growth_gabe

They let me go pretty deep. I mocked up a few mental models for what friction looks like (cell order dependency, lack of Git-native workflows, debugging UX in notebooks). The interviewer had opinions and pushed back, which made it feel like a real product conversation. I liked it.