Round 1 was ML design, a recommendation system. No problems, it was a relaxed conversation and I got through it easily.
Round 2 was a project deep dive. Over the past few years I moved from ranking to LLM agentic work, so I prepared LLM-related projects. But the two interviewers who showed up both had ranking backgrounds and were probably not very familiar with agentic projects. In the end their feedback was that the main work was data & harness, and that there was no ML depth signal.
Round 3 was the hiring manager behavioral round. It was average, no problems.
My takeaway is to prepare two project deep dives and pick whichever fits the situation. I did not want to prepare a ranking project because I have not worked on ranking in the last few years, the tech stack from back then may be outdated, and I have probably forgotten some of the details.
Discussion
Loading comments…