Harmattan AI interview questions

Four rounds, they hire DEs in Paris but are flexible to Lausanne, CH. Each interview is followed up in a few days to 1 week max. Overall good but hadn’t prepared for MLOPs as was told it was a standard DE systems interview, thus rejected after round 3, definitely brush up on MLOPs. Round 1: HR Screen (30 mins) Talent Acquisition reaches out to get to know you, your interest for joining, skillset, and what the company is working towards. Round 2: Coding (60 mins) Use either Python or C++ in CoderPad. Standard DSA questions mainly around Binary Search Trees and Strings. Ask questions and clarify any assumptions. Round 3: Systems Design (90mins) Interview with two engineers (1 DE, 1 MLEng) to discuss your background and projects (30 mins) and then remaining 60 mins for systems interview. Make sure to know data ingestion in-depth and tech stack, and more importantly, MLOPs as the questions are heavily around being able to create data platforms for ML continous training (MLOPs). Make sure to clarify any FR & NFR requirements and take into account storage and performance costs/tradeoffs. Round 4: Coding + Systems Final round Similar to rounds 2 & 3 but now as an “onsite” interview. Usually this is remote unless you are in Paris (main HQ).

I had a very disappointing interview experience with this startup. The process involved too many phases and repeated conversations, but there was never any real clarity about which role I was being considered for. It felt like they were trying to figure out where I might fit, while not even being sure themselves what they actually needed. With 20 years of experience in the field, I expected a more structured and professional process. Instead, it came across as disorganized and uncertain. In the end, it took me just one interview elsewhere to secure a better position with a large, established organization. That contrast made it even clearer to me that this company did not seem to know what it was doing. Overall, it felt like a typical startup mess.