Looking for ML roles in Europe took longer and was more tiring than I expected. I sent 32 applications over about three months, and the rejections followed a clear pattern: most teams either required a PhD or wanted only experience with LLM fine-tuning.
My background is recommender systems and production ML tooling, the unglamorous work that keeps models running at scale, and that combination fitted badly with job descriptions written with transformer research in mind. I reached interviews at six places, which felt low for the number of applications I sent, and some of those early rounds ended with feedback that confirmed I was not what they were looking for. The offer came through a referral. A former colleague from a previous startup had joined Zalando in Berlin and gave my name internally when a Senior ML Engineer position opened. That seemed to make people read my application differently. The recruiter wrote to me within a week, which had not happened anywhere else in this search. The process was rigorous: three technical rounds about system design, coding, and my recommender work in production, plus one more interview that surprised me. They called it a paper-discussion round. I had to argue both sides of a recent ML paper ↓