Rokt Senior Full-Stack Machine Learning Engineer Onsite Interview Experience

This post was last edited by Anonymous on 2025-10-08 at 09:27. The job requirements are a full-stack senior MLE (Model-View-Language) role, roughly 60% ML and 40% infrastructure. The interviewer said

This post was last edited by Anonymous on 2025-10-08 at 09:27. The job requirements are a full-stack senior MLE (Model-View-Language) role, roughly 60% ML and 40% infrastructure. The interviewer said it was ML fundamentals, but directly asked me to design video recommendation systems for platforms like Instagram and TikTok, emphasizing model and data labeling. I was very familiar with these tasks. The American female interviewer said, "genuinely happy to meet you," and the recruiter said they had never seen such good feedback. A bit flattered. VO: 1. Business Qualifications (BQ) Director Interview, generally speaking, the feedback was good. 2. Coding question: Determining if abbreviations match the original words. Slightly harder than coding challenges, but not too difficult overall. I overcomplicated one part and got stuck. The follow-up wasn't fully tested in time, but I basically finished. 3. The most comprehensive system design ever, e-commerce recommendation. The first 25 minutes focused on machine learning (ML), the last 25 minutes on ML infrastructure (how to design the entire pipeline, how to store features, and how to optimize). Basically, a comprehensive understanding of ML is required. The interesting part is that the feedback was "very strong candidate," and the skills were very aligned, but I still didn't pass. Quite ridiculous. I guess it's because the coding wasn't perfect, or because I haven't done much ML in the past two or three years due to focusing on ML infrastructure...