Design a product recommendation system.

"The process was much quicker and smoother than at other companies. Recruiters shared some pointers on what to prepare, and mostly it was around that only. What did not go well was that one interview was online, and others were on-site."

"I interviewed for an M1 Engineering Manager role, and the process included a screening round first, followed by the main loop. The screening consisted of one broad behavioral interview and one system design interview, in which I could choose between machine learning systems, software systems, and products; I chose machine learning. In the main loop, I had another ML system design round, a project deep dive, a team management interview, another behavioral round, and an AI-assisted coding interview. The AI-assisted coding piece was the newest part of the process for me, but the bigger surprise was how behaviorally heavy the EM loop was. My main takeaway is that they expect you to provide depth on your own, even if the interviewer isn't especially good at eliciting it."

"I interviewed for an M1 Engineering Manager role, and the process included a screening round first, followed by the main loop. The screening consisted of one broad behavioral interview and one system design interview, in which I could choose between machine learning systems, software systems, and products; I chose machine learning. In the main loop, I had another ML system design round, a project deep dive, a team management interview, another behavioral round, and an AI-assisted coding interview. The AI-assisted coding piece was the newest part of the process for me, but the bigger surprise was how behaviorally heavy the EM loop was. My main takeaway is that they expect you to provide depth on your own, even if the interviewer isn't especially good at eliciting it."

"I interviewed for a Senior Staff MLE role, and the biggest thing I’d say is Reddit felt pretty team dependent, so I wouldn’t assume every team runs the exact same loop. Mine started with a recruiter screen that surprisingly included a couple very basic ML questions, then a hands-on technical screen where I had to work with data in a notebook-style environment and get a model actually running. The final loop was ML fundamentals, ML system design, coding, and a hiring manager round. The ML system design round was the most interesting because I had prepped for a modeling-heavy conversation and got pushed much more into ML infra, like deployment, latency, feature pipelines, and how the A/B test would work under the hood."