Data Scientist Interview Experiences

Same OA format as FDSE I think but onsite was stats focused, hypothesis testing and model evaluation metrics. Behavioral was embedded into every round, that was different. Coding bar felt pretty high.

Online assessments took forever verbal numerical then the games then HireVue after and final round two interviews where partner brought up influencing a team and senior manager wanted to know about failure. standard but conversational

A/B testing setup and how to handle p hacking and peeking at results came up and I wasn't expecting that at all. First technical round was statistics and probability, questions about distributions and how you'd design an experiment with finite sample size constraints. Second round was a machine learning system design question about building a merchant recommendation engine at Shopify scale. Also a values interview that felt more human than most. Long process overall but the interviewers were smart and treated it like a real technical conversation rather than a gatekeeping exercise

Prescreen with campus recruitment was chill then assessment day was interesting, they put us in groups to work through business cases and present to a panel with a debrief after each one. Final round was just about work experience and industry knowledge, I think the group exercises are where they actually evaluate you

Recruiter screen then technical assessment medium LC then two 45 min interviews with the managers. Process took 6 weeks, felt long but the actual interviews were reasonable

45 min on Teams they went through my resume and wanted to know about a REST API I built and leadership questions too, felt casual but took over a month to hear back

Online assessment first with situational judgment and pattern recognition and logical reasoning stuff. If you pass that you get a 1 on 1 interview with 4 behavioral and 2 situational questions. Then the panel, 6-7 behavioral questions in one hour using the CAR format not STAR. A lot of questions one after another in a short window

Typical behavioral questions and then some technical questions like writing sql query, data cleaning, and then a case study on machine learning techniques. Interviewers were very friendly and accomodating

Technical phone screen had SQL and a stats question on modeling surge pricing demand. Explained the approach fine but yeah

CUDA and GPU memory questions which I was not prepared for as a DS candidate. also Python and ML model selection. Nvidia DS is very different from other companies because of the GPU focus

Multi-stage interview First round was with HR, I was asked standard behavioural questions about myself and my interest in the role. Second round was a technical interview. Had to complete 2 easy/medium array manipulation-type coding questions on the spot and walk through my approach. Next I was asked some theoretical questions relating to statistics and data manipulation (topics included variance, confusion matrix ,etc). The last question was a logic puzzle about pouring water in a cup

OA had two LC medium questions, plus work simulation questions, then final round was heavy on leadership principles, brought up disagreeing with someone and they push for specific detail on your examples

They care a lot about how you explain your approach and the tradeoffs you made. Recruiter screen then technical, OOP modelling of financial data, then a project showcase where I presented my best ML project. Good feedback but I didn't make the cut in a competitive cycle

Phone screen with recruiter then another call with ops team, then take home assignment analyzing some operations problem. Haven't done the final round yet but people have been nice and the take home was interesting

They wanted my research background summarized in 90 seconds for the pre recorded video, plus two other timed questions. Multiple rounds after, behavioral and one technical on regression modeling and handling missing data. Waiting on final round feedback now

Live coding with 2 interviewers fizz buzz type question then mostly situational based behaviourals like about leadership etc

The HackerRank OA had three parts, coding then SQL then an API task and I thought I did fine on coding but the SQL was rough

Phone screen with someone from data asked about experiment design and SQL. Technical round was a case on measuring delivery ETA accuracy.

Everything happened so fast, OA 50 MCQs in 35 minutes then same day got called for DSA round with hashset and tree questions. Second DSA next day was harder DP on subsequences, manager got into a guesstimate about daily deliveries in Bangalore then rapid fire about resume claims process was 2 days total.

Asked about my resume in depth- past projects and how they connected to what I am studying

HR screen was fine then the technical round had JavaScript React and C# questions, weird combo but their stack uses all three, LeetCode easy level questions and some SQL plus OOP concepts, took 3 weeks total which they warned me.

First half was behavioral then asked some easy technical questions related to unsupervised learning and difference between overfitting/underfitting. They were friendly and one interview only before I got the offer

application took forever just to set up the account. Got an email for an interview like 3 months later and by that point I forgot I even applied. Two people on the call asking basic behavioral stuff, tell me about a time you worked on a team, how do you handle deadlines. Also some questions about databases which were easy. It was painfully slow and I already accepted something else by the time they got back to me

C# came up briefly even though the role was data science focused, just awareness questions not coding. Technical had OOP questions and then a short problem involving string manipulation in Python. SQL query too

15 mins, really easy, heard it's way harder for FTEs with like 4 rounds but for intern ig they didn't really care all that much

Took 1 week total 2 interviews a couple days apart. Standard process HR round then manager round, technical questions were focused around CI/CD and Github mostly