Ten firsthand reports on Mastercard Data Science and AI/ML internship interviews from 2022 to 2026.

Coverage: Data Science, Data Scientist, and clearly related AI Garage internship roles. The reports cover India, Lisbon, Gurgaon, and one location that was not shared. They include online tests, recorded video screens, technical interviews, project deep dives, and HR rounds.

For a fren, thank them for their patience. And good luck to them!

The short version

There is no single Mastercard interview loop for this role.

Across the ten reports, candidates mentioned more than 40 interview questions, tasks, or topic prompts. They were all reported as part of an OA or interview. This does not mean 40 complete practice problems: some were exact questions, some were only topic names, and some were unnamed coding sections.

1. Your projects are the safest place to start. Nine of the ten candidates were asked about their resume, projects, or research. Know what you built, why you made each choice, what went wrong, and what you personally did.

2. Coding appears often, but not in every interview. Five reports included coding or DSA. One candidate was asked no DSA at all. The only exact Intern problem we could identify was Container With Most Water.

3. The ML questions were practical and basic more often than exotic. Candidates saw cross-entropy, class imbalance, regularization, CNNs, clustering, CatBoost, probability, and model choice. Several questions started from the candidate's own project.

4. Some India campus loops were much longer. They included timed tests, DSA, puzzles, SQL, ML, project rounds, and HR. The Lisbon and 2026 reports described shorter or more team-specific processes.

5. SQL matters, but the evidence does not show it in every loop. One detailed campus report included join types and live join exercises. The other nine reports did not give a clear SQL question.

6. Behavioral and product questions are part of the job. Candidates were asked Why Mastercard, how they work with others, when they used data, and how they would improve a product.

A practical way to prepare: learn your projects cold, review core ML and probability, practice basic arrays, strings, trees, graphs, and SQL joins, then prepare short answers for Why Mastercard and teamwork.

Online tests and early screens

OA 01: A fast three-part campus test

Campus cycle: 2025–26 · Role family: Data Science / AI-ML · Process: IIT Delhi

1. Twenty aptitude questions in 12 minutes.
2. A separate ML and deep learning section in 12 minutes.
3. Two DSA questions in 40 minutes.

The exact DSA questions were not included in the report. After the test, the candidate had two technical rounds and an HR round.

OA 02: Two coding problems in two hours

Interview cycle: 2023 · Role: Data Science Intern · Process: India campus

The candidate solved two basic string and array problems. They described the level as roughly beginner-to-intermediate competitive programming. The exact statements were not shared.

The later rounds covered probability puzzles, basic ML, SQL joins, a cab-fare project, and HR.

OA 03: ML, probability, statistics, and one coding task

Interview month: November 2022 · Role: Data Scientist Intern · Location: Gurgaon

The assessment covered machine learning, deep learning, probability, statistics, and one coding problem. The candidate did not share the coding prompt.

The live interview then moved to resume projects, ML, and CatBoost.

Screen 04: Recorded video before two final interviews

Interview month: January 2026 · Role: Data Scientist Intern · Location: Lisbon

The candidate sent a resume and cover letter, completed a recorded video screen, and then had two final interviews. The whole process took about three weeks.

The clearest question was simple but open-ended: choose a favorite product, explain why you like it, and name two things you would improve.

Reported interview topics and tasks

Coding

1. Solve two basic string and array problems in two hours. The exact tasks were not shared.
2. Solve two DSA questions in 40 minutes. The exact tasks were not shared.
3. Find the maximum water held between two pillars. This matches Container With Most Water. The original report did not include the full array model, so the practice page states that missing detail clearly.
4. Complete one coding task after an ML, deep learning, probability, and statistics test. The prompt was not shared.
5. Complete a coding section during a technical interview. The prompt was not shared.

That is the honest limit of the coding evidence. We did not turn broad labels such as graphs, DP, trees, or “two array questions” into made-up problems.

SQL, probability, and puzzles

1. Explain the main SQL join types.
2. Solve join exercises during the interview. The tables and queries were not preserved.
3. Work through a burning-rope puzzle.
4. Solve a socks puzzle.
5. Answer a probability question involving dice.
6. Handle an unnamed probability question that the candidate compared with a classic interview puzzle book.

Machine learning and statistics

1. Explain cross-entropy loss.
2. Explain how you would handle class imbalance.
3. Choose an ML method for a given situation and defend the choice.
4. Explain regularization.
5. Explain how a CNN works.
6. Explain CatBoost and compare it with gradient boosting.
7. Explain clustering and when you would use it.
8. Explain the architecture used in one of your projects and why you chose it.
9. Show how you would solve a practical problem with data science.
10. Answer general ML, deep learning, and probability questions tied to your resume.

Projects, research, and behavior

1. Walk through a cab-fare estimation project in detail.
2. Explain a graph-learning project about imbalanced node classification.
3. Explain the goal, data, architecture, implementation, and result of a resume project.
4. Be clear about your own contribution, not only what the team built.
5. Why Mastercard?
6. Tell me about a time you worked with data.
7. How do you work with other people?
8. Tell me about a teamwork experience.
9. What could you bring to the team?
10. Why are you interested in data science?
11. Pick a favorite product and suggest two improvements.

10 candidate timelines

01: One behavioral interview and one difficult technical interview

Report date: February 25, 2026 · Interview month: January 2026 · Role: Data Scientist Intern · Result: No offer

The candidate applied online and finished the process in about one week.

The behavioral round covered how they work, Why Mastercard, and a time they used data. The technical round included an unnamed coding section and resume questions. The candidate found the behavioral interview easy and the technical interview hard.

02: Lisbon video screen and two final interviews

Report date: February 5, 2026 · Interview month: January 2026 · Role: Data Scientist Intern · Location: Lisbon · Result: No offer

The process started with a resume and cover letter. A recorded video screen came next, followed by two final interviews.

The report preserved one product question: choose your favorite product, explain why, and propose two improvements. It is a good reminder that a data science interview may test product judgment as well as technical knowledge.

03: IIT Delhi test, two technical rounds, and HR

Campus cycle: 2025–26 · Role family: Data Science / AI-ML · Process: IIT Delhi

The test had three timed parts: 20 aptitude questions in 12 minutes, ML and deep learning questions in 12 minutes, and two DSA problems in 40 minutes.

The first technical round focused on the resume, projects, ML, and deep learning. The second included an unnamed brain teaser, an ML scenario, and a deeper discussion of complex projects. HR covered personality, Why Mastercard, career goals, and interests.

The source did not preserve the exact DSA, brain-teaser, or ML-case prompts.

04: Gurgaon DSA, ML fundamentals, and one project

Report date: October 27, 2025 · Interview month: March 2025 · Role: Data Scientist Intern · Location: Gurgaon · Result: No offer

This was the most balanced technical report in the set. The candidate was asked to solve the maximum-water-between-pillars problem, explain cross-entropy, discuss class imbalance, choose techniques for an ML situation, and explain a resume project.

The coding task is the one exact Intern question we can link: Container With Most Water.

05: Data science terms, clustering, and a practical case

Report date: August 3, 2025 · Role family: Mastercard AI Garage internship · Process: IIT Bombay

The first round started with the resume. The interviewer picked data science terms from it and asked for clear definitions, including clustering.

The second round asked the candidate to explain how they would solve a practical problem with data science. A teamwork question followed. The candidate later worked on imbalanced node classification in graphs during the internship.

The public page did not print a formal HR title, so this roundup keeps the broader AI Garage role-family label.

06: ML projects, architectures, and probability

Report date: August 3, 2025 · Role family: Mastercard AI Garage AI/ML internship · Process: IIT Bombay

The interview stayed close to the candidate's work. They discussed ML projects, the architectures they used, and probability.

The candidate also said a harder probability puzzle could appear late in the process, but the exact question was not shared. The later internship work involved graph neural network research.

07: Two rounds built almost entirely around projects

Report date: July 21, 2024 · Role: Data Scientist Intern · Internship location: Gurgaon · Result: Selected

The candidate had two interviews, and the second was short. Most questions came from the resume: explain the projects, explain the ML architectures, defend those choices, and say why data science interests you.

This candidate was not asked generic ML questions, puzzles, or DSA. That directly conflicts with the coding-heavy campus loops, which is why one fixed Mastercard interview template would be misleading.

08: Basic ML questions followed by project discussion

Report date: April 30, 2024 · Role: Data Science Intern · Result: No offer

Most of the interview covered basic ML algorithms and projects. The clearest questions were: What is regularization? How does a CNN work? Explain the projects on your resume.

The candidate felt they made the answers too complicated. Their own advice was to begin with a simple, correct explanation and add detail only when asked.

09: Coding, puzzles, SQL, and a cab-fare project

Interview cycle: 2023 · Role: Data Science Intern · Process: India campus · Result: Selected

The first stage was a two-hour test with two basic string and array problems.

The first technical round lasted about 45 minutes. It covered burning-rope, socks, and dice puzzles, basic ML, SQL join types, and join exercises.

The second technical round lasted about 30 minutes. The candidate explained a cab-fare estimation project, what they knew about Mastercard, and what they could bring to the team. HR then confirmed the selection, stipend, and location.

10: Gurgaon test followed by CatBoost and project questions

Report date: October 31, 2023 · Interview month: November 2022 · Role: Data Scientist Intern · Location: Gurgaon · Result: No offer

The process began with a test on ML, deep learning, probability, statistics, and one coding problem.

The interview focused on resume projects and ML. The clearest questions were: What is CatBoost? How is it different from standard gradient boosting?

The coding prompt was not shared, so it should not be guessed.

What repeated often enough to prioritize

1: Know your own work

Nine of the ten reports touched the resume, projects, or research. Prepare one project from start to finish: the problem, data, baseline, features, model choice, metrics, mistakes, result, and your contribution.

2: Review the core technical tools

Be comfortable with basic ML, probability, Python or DSA, and SQL joins. For ML, start with a short correct answer. Then explain the tradeoff or example if the interviewer wants more.

3: Prepare for the team in front of you

An AI Garage research team may go deep on architectures, papers, and graph learning. Another team may care more about a product question or a short behavioral round. Read the job description and adjust your examples.

P.S. I also found Special Sum of an Array in a Data Scientist fresher report, but that was not an internship, so it is not counted in the ten experiences above.