Microsoft Data Scientist Interview Guide: OA, Back-to-Back Rounds and Experiment Design
Prepare for Microsoft data scientist interviews with reported assessment patterns, interview questions, and preparation advice.
On this page
Last reviewed November 15, 2026.
For data scientist interns, new grads and experienced candidates.
Built by the Codinzhub team from Microsoft data science interview experiences that candidates shared, including the questions posted on Codinzhub. Each experience is one person's interview, so use this as a map, not a promise.
The 30-second version. A online coding platform OA (one coding problem and one data-processing problem, or Python and pandas tasks for campus intern roles), then three back-to-back interviews on one day. Rounds mix resume deep dives, ML fundamentals, a coding or SQL question and A/B testing. Senior loops added an AI-assisted ML round in your own environment. Processes can be slow: one took months from applying to a decision.
Round-by-round breakdown
Stage 1: Application, where candidates applied
How it works. Experienced candidates mostly cold-applied; one waited nine weeks for a recruiter. Campus data science intern OAs in 2026 were held at IIT BHU and IIT (ISM) Dhanbad. One candidate advised using an Outlook email, because Gmail replies kept getting lost.
Stage 2: Shortlisting, how candidates moved forward
How it works. The OA comes first. An interview invite followed three days to three weeks later. One senior candidate had a short phone screen on the resume and online experimentation instead.
Stage 3: Online assessment
At a glance
- Length: about an hour
- Questions per round: 2 questions, or a mix of DSA and Python data tasks
- What was asked: one coding problem, data processing with pandas, NumPy and SciPy
- Most common: a pandas analysis task
Microsoft DS OA questions shared on Codinzhub:
| Question | Drive | Topic |
|---|---|---|
| Employee Sales Analysis Using Pandas | Campus | Pandas |
| Top Employee Across Branches | IIT BHU | Pandas |
| Numerical Optimization Using SciPy | IIT (ISM) Dhanbad | SciPy |
| Minimum Operations to Equalize Prices | IIT BHU | Arrays |
| Count Strictly Increasing Subsequences of Length K | Campus | DP |
| Count Palindromic Substrings | Campus | Strings |
| Maximum XOR with Equal Bit Length | IIT (ISM) Dhanbad | Bits |
More questions: Browse Microsoft questions.
Stage 4: The interview rounds
How it works. Three interviews back to back, about three hours in total, often with senior people. Each round mixes a resume deep dive with one technical area.
Round 1: Resume, behavioral and ML fundamentals
At a glance
- Length: about 45 minutes
- Questions per round: behavioral questions plus 4 to 6 ML questions
- What was asked: your research, conflict with teammates, KNN and other ML basics
- Most common: questions tied back to your resume projects
Questions candidates reported: walk through your research projects; a conflict with a teammate; ML fundamentals such as KNN; rapid-fire LLM questions for candidates with LLM research.
Round 2: Coding or SQL
At a glance
- Length: 20 to 45 minutes of coding after a resume discussion
- Questions per round: 1 coding or SQL question
- What was asked: trees and BFS, SQL, plotting metrics
- Most common: no single question repeats
Questions candidates reported:
| Question (our summary) | What it tests |
|---|---|
| Print an org chart level by level | Tree, BFS with a queue |
| Draw a precision-recall curve from scratch | Metrics, coding |
| A SQL question, then A/B testing | SQL |
More questions: Browse Microsoft questions.
Round 3: Experiments, cases and the hiring manager
At a glance
- Length: 45 to 60 minutes
- Questions per round: 1 or 2 business cases
- What was asked: A/B testing, what to do without an A/B test, Type I and II errors, metric design
- Most common: experiment design for a product change
Questions candidates reported:
- What to do if you cannot run an A/B test: ways to estimate the counterfactual.
- Confusion matrix and Type I and Type II errors.
- Design an experiment for adding a product option, and interpret the results.
- Analyse pilot users of an AI product.
- A business scenario the team is working on: how you would judge each option.
More questions: Browse Microsoft questions.
AI-assisted ML round (senior)
How it works. You get a CSV and build a model in your own local environment, with AI tools allowed and encouraged. The candidate who reported it had a new laptop with no setup and could not finish. What is judged is how you direct the tool and check its output.
Stage 5: Decision and offer
How it works. Offers came about a month after the interviews for two candidates. One senior candidate waited five weeks and was rejected on level calibration despite positive rounds, then got feedback by asking politely.
The Codinzhub Next-Round Playbook
What data scientists who moved on did. Put together by the Codinzhub team from the DS experiences in this guide.
To get past the OA
- Practise pandas and NumPy tasks under time pressure, plus one medium coding problem.
To get past the interviews
- Know every project on your resume in depth. Almost every question came back to it.
- Prepare ways to evaluate a change without an A/B test.
- Revise ML basics and, if relevant, LLM fundamentals.
For senior roles
- Set up your local Python environment and an AI coding tool before the ML round.
- Write your resume at the level you want. One candidate passed the senior screen but not the DS II screen with the same resume.
About this guide
This guide is researched and written by the Codinzhub team. We read every Microsoft data science interview experience we could find, kept only what candidates actually described, and rewrote it in our own words. Questions are our own short summaries. Found something out of date? Share your own experience on Codinzhub and we will update the guide.
Keep track of what you’ve learned.