Dataiku Interview Questions
& Process

The interview process was professional and organized. However, I received conflicting feedback. They initially praised my technical skills in data science and ML models, but then stated the role requires less technical work. If you have the technical background listed in the job description, be prepared for the possibility of being told you are overqualified. Overall, it's a great company with a good product, but this specific role seems more focused on client management than actual customer success. Candidates with technical backgrounds might want to reconsider unless they are comfortable with a role that is approximately 90% account management and 10% technical work, where their technical knowledge would primarily be used for delegation.

The interview process started with an HR interview that went very well because I had researched the company. Then, a technical interview went very well, with two general questions to propose solutions for problems like counting letter occurrences in a sentence, while considering the solution's complexity. Next was a technical test, which was hard but I passed it, asking about 4 or 5 times the approximate duration the employer expected. This was followed by an operational interview, with feedback on the technical test, questions about my background, and discussions in both English and French. I didn't pass this stage because I didn't explain my technical test solution well and didn't highlight my past projects effectively. The final stage was a friendly HR feedback interview.

So the whole thing started with a chat with a recruiter, just to get a feel for the company, what they do, and what roles are available. I also gave them a rundown of my education and work history. Next up was a technical interview where I had to solve a coding problem. The last part was a take-home assignment. The task was to build an API that finds the shortest path between two spots, making sure to dodge any obstacles. They were looking for solutions using algorithms like Dijkstra’s or maybe A*. It wasn't just about getting the right answer and making it fast, but also about having good code quality, which meant writing clean, easy-to-maintain code, adding unit tests, having logs and secure headers, and generally submitting something that’s pretty much ready for production.

I had an initial chat with a technical recruiter, then spoke with two SDETs for a technical discussion. After that, there was a take-home technical test that I had a week to complete. The final step was an interview with the VP of Test. The first two parts went really well; the recruiter was great, and the chat with the SDETs was good too, covering some technical points. However, the take-home test was a major letdown, not what I'd expect from a leading data science company. The task required resetting the database after each API test run, which is technically flawed and causes concurrency issues. I explained this in detail during my meeting with the test lead, but there seemed to be a disconnect between expectations and best practices. He listened but didn't really acknowledge the issues, just mentioned a potential post-mortem. This reset approach is unheard of, even in places like Dataiku. I also pointed out that validating data via API tests isn't their job; that's for integration tests, and the exercise ignored the testing pyramid, leading to inefficient validation. They seemed to struggle with running a very simple project despite clear instructions, which is worrying for a tech company. Overall, it was disappointing. The effort for the test didn't match the feedback or expectations. They need to fix these technical issues in their tests or risk losing good candidates.