Uber Senior Product Manager Interview Experience
Uber · Senior · Product Manager
Rehearsing your take-home top-down isn't actually that helpful. There were so many curveballs thrown in the middle of the jam session that it would be almost impossible to just read a script and go slide to slide — you need to be on the tip of your toes.
Interview process
The process was a short recruiter screen, then a hiring manager round that spent way more time on an older role on my resume than on my current job, plus a rushed mini jam at the end. After that I had to do a take-home deck for an events marketplace prompt and present it in a one-hour jam panel, then go through three more final interviews spread across three days.
The loop felt inconsistent because nobody besides the hiring manager was actually from the team, and each interviewer had a completely different style, from rapid-fire generic PM questions to aggressive marketplace probing to a bizarre senior director case. I got rejected, but Uber did send unusually detailed written feedback afterward, which almost made it more frustrating because the strengths section was so strong and the main ding was marketplace economics that felt very insider-specific.
Interview rounds · 4
- 1
Recruiter screen
BehavioralI got an inbound recruiter call that was really short and mostly informational. It felt more like they were selling me on Uber than screening me hard.
Q1. Can you walk me through your background, comp expectations, and interest in the role?
How they answeredI kept this one very high level. I gave a quick summary of my background, my salary expectations, and why my logistics experience overlapped with the role. Since they had reached out to me, it did not feel intense at all. The recruiter even spent time talking about how much better Uber was than my current company, so the vibe was more pitch than screen.
- 2
Phone screen
BehavioralProduct DesignExecutionMy hiring manager screen was an hour and felt way less structured than I expected. He spent most of the time digging into an older logistics role on my resume, then used the last 15 to 20 minutes for a mini jam session that got stressful fast.
Q1. In your last-mile role, how did you scale the product to new countries?
How they answeredHe went deep on a role I had done almost three years earlier and treated it like it was my current job. I talked through scaling to new countries, adoption issues, and how I worked through legal and compliance constraints in different markets. The hard part was remembering exact outcomes and numbers on the spot, so a couple times I gave ballpark savings in hours or money instead of precise figures. I did not want to hide behind saying I just did not remember.
Follow-up questions- Did you run into adoption challenges in other countries?
- Did you have challenges with legal counterparts?
- What would you do in a market like Germany with heavy regulatory compliance?
Q2. What would you launch for riders to signal when they feel uncomfortable in an Uber?
How they answeredI started the way I had prepped, with clarifying questions and a framework, and he cut me off after a couple minutes and basically said to get to the juice. We narrowed it to a real safety situation and I suggested an in-app button that could alert Uber support that there was a problem with the driver and trigger follow-up or coaching. He immediately pushed on the gap between post-ride support and in-the-moment safety, especially if the rider needed to get out now. I was pretty flustered because the interview suddenly stopped rewarding the structure I had prepared.
Follow-up questions- Define what uncomfortable means.
- Is it too hot or too cold, or is it a real safety issue?
- If the situation is a single woman feeling uncomfortable in the car with a male driver, what would you build?
- What if there are still 30 minutes left in the ride and her safety is at risk right now?
- 3
Take-home assignment
Case StudyPresentationProduct StrategyExecutionAfter the hiring manager round, they moved me forward the next business day and sent me the jam prompt plus prep docs. I had about two weeks to make a slide deck and had to send it 24 hours before the interview.
Q1. How would you make Uber to and from large events with 10,000+ attendees a magical experience for riders and drivers?
How they answeredI built my deck around dedicated pickup zones near events, credits to nudge riders to walk there, and an airport-style FIFO queue for drivers. I laid it out as a phased MVP and talked about using pilot programs where historical data was limited. I also used AI to make some mocks so I had something concrete to point to. I practiced by recording myself, but honestly the real lesson was not to over-script because the actual jam was way more interruption-heavy than a clean presentation.
Follow-up questions- What are the biggest problems, and do they differ by event type?
- What marketplace effects matter here?
- What does the magical UX look like?
- How would you measure success?
- What tools or systems would support it?
- 4
Final / onsite round
PresentationBehavioralExecutionTechnicalAnalyticalProduct StrategyThe final round was spread across three days and felt like three straight days of getting grilled. It started with the one-hour jam panel, then a PM round, then a technical round and a senior product leadership round. What stood out was that none of these people were from the actual team, so every interviewer brought a totally different style and set of assumptions.
Q1. Why did you choose dedicated pickup zones for the event experience, and why not a different solution?
How they answeredFrom my title slide they were already interrupting, so it turned into a rapid-fire defense instead of a presentation. I explained the dedicated pickup zones, rider credits for walking there, and the FIFO driver queue as a way to reduce cancellations and chaos. The toughest pushback kept coming back to driver earnings and whether I was hurting drivers with extra wait time or constraints. I tried to talk through tradeoffs across riders, drivers, and the support center, but it definitely felt adversarial and pretty brutal.
Follow-up questions- How would this affect driver earnings?
- Are you penalizing drivers with this flow?
- How would support handle this operationally?
- What other marketplace problems could this create across riders, drivers, and support?
Q2. How do you work cross-functionally, unblock engineering, and prioritize features?
How they answeredThis round was basically generic PM essentials asked one after another. I answered with examples about cross-functional alignment, unblocking engineering, and how I prioritize, and I tried to inject data and outcomes even when the questions were broad. The weird part was there were almost no follow-ups at all. It was just me answering, hearing 'okay,' and then getting the next question, so it felt more like surviving a queue than having a real conversation.
Q3. How do you work with data scientists and engineers?
How they answeredI thought this part actually went well as I frequently partner with data science and engineering already.
Q4. How do you balance tech debt with feature priorities?
How they answeredI walked through how I think about balancing new features against tech debt, using examples of how I trade off speed, risk, and longer-term platform health. It felt straightforward in the room, which is why I was surprised later that I still got flagged on technical experimentation nuance.
Q5. Uber is launching autonomous vehicles alongside human-driven cars. How would you decide whether to launch, and what would your success metrics be?
How they answeredHe wanted an experimentation framework more than a product idea. I talked through launch criteria, success metrics, and how I would compare autonomous and human-driven experiences, then he kept drilling into A/B testing and cohort design from a data science angle. I was not really expecting that level of experimentation depth from the prompt, so I had to pivot into methodology mode pretty quickly. In hindsight I should have gone deeper on experimental design earlier instead of treating it like a normal metrics question.
Follow-up questions- How would you structure the experiment?
- What A/B tests would you run?
- How would you use cohort analysis?
Q6. In the Netherlands, if couriers are employees instead of contractors and you cannot require certain hours or penalize missed shifts, how would you improve meeting the delivery promise window?
How they answeredThis one completely threw me. I understood it as a regulatory and incentives problem where supply was not matching demand, but you also could not force behavior the usual way. I started brainstorming things like in-app guidance, rewards, and achievement-style incentives to nudge people into taking more shifts or blocks, but I was fumbling for most of it. The interviewer was very senior, pretty cocky, and I could tell he knew I was struggling, which made the whole thing feel even more unhinged.
Follow-up questions- Would you use in-app guidance?
- Would you add an achievement bar or rewards?
- How much would you pay people to change behavior?
Tips from the candidate
Do not only prep stories from your current job. They went straight to the most relevant experience on my resume even though it was old, and I had to reconstruct results live. For the jam, know your framework but do not cling to it too hard, because if you stay in clarifiers too long they may cut you off and ask for the actual solution. If the prompt smells like marketplace, be ready for repeated pushback on driver incentives, earnings, ops, and support consequences. Also prep experimentation more deeply than just naming metrics. Think pilots, A/B tests, cohorts, and explicit launch criteria.
Company culture
Outside of the hiring manager, I did not meet with the actual team, so the loop felt like a generic Uber PM loop more than a targeted team match. The interview prep docs and labels did not line up that well with what I actually got. Several rounds were much more probe-heavy and adversarial than advertised. The one positive was that they did give real written feedback afterward, which is rare.