Thomson Reuters Data Scientist candidates describe a mix of ML and Python concept checks, project deep-dives, and a behavioral finale. Formats vary by team, so prepare for both timed coding and conversation.
$126K
Avg. Base Comp
$185K
Avg. Total Comp
3 rounds
Typical Rounds
Not reported
Process Length
Candidates describe Thomson Reuters Data Scientist interviews as more conceptual than algorithmic, though the format clearly varies by team and location. One candidate reported a casual background chat, then a skills round on pandas, numpy, precision and recall, data types and data ethics, then a relaxed STAR-style behavioral round. That candidate received an offer and found the conceptual material harder than the coding.
Another candidate who received an offer described a HackerRank assessment of 30 questions in 45 minutes. A technical round on ML and Python followed. The final round was with about three managers, who explained the mission and asked about customer interactions and how the candidate organizes their work.
Not every experience was smooth. A candidate for a Bangalore-based role was told to expect coding but faced a roughly 30-minute ML theory session covering multicollinearity, tree-based models, bagging and boosting, F1 and imbalanced data. Another candidate completed an NLP and multiclass classification homework and a combined technical and HR conversation, then heard nothing back. A different candidate reported a research talk during which a senior manager made pointed comments, plus a basic linked-list coding question.
Practice the fundamentals: bias-variance, evaluation metrics, imbalanced data and cross-validation. Be ready to walk through your own projects and explain why you made each design choice. Ask the recruiter what the technical round covers, since reports about format conflicted. Practice speed on easy coding problems as well.
Synthesized from 6 candidate reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Thomson Reuters process.
The interview started with a first round that was mostly introductory—the hiring manager walked me through the team's work and what the role would look like. Then came a long wait before the second round, which arrived around a month and a half later. That second round was a homework assignment centered on NLP and multiclass classification. During the technical and HR combined interview, they asked me to dive deep into my project experience and tested conceptual knowledge. The questions felt approachable—basic to medium level—with a lot of focus on walking through what I'd actually built and why. I found the experience overall straightforward, though I was aware that communication mattered during the interview, and I felt a bit of nervousness that affected how well I could articulate my thoughts, even when I knew the material. After completing both rounds, everything went quiet. I never heard back on the outcome, and it seemed like the company just stopped communicating. That silence was frustrating—I had completed the full process but got no closure.
Prep tip from this candidate
Have concrete examples of your past projects ready to discuss in depth. The NLP and multiclass classification problem came up, so if you're interviewing for a similar role, review those concepts. Most importantly, practice articulating your answers clearly—communication seemed to carry real weight, and the long silence after rounds made it hard to know if technical depth or interview delivery was the deciding factor.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Thomson Reuters
The role of A/B testing in measuring the success rate of an analytics experiment
| Question | |
|---|---|
| Fine-Tuning VS RAG | |
| Bias vs. Variance Tradeoff | |
| Data Preparation for Imbalanced Data | |
| Boosting Instagram Stories | |
| Your Strengths and Weaknesses | |
| Docs Metrics | |
| WhatsApp Metrics | |
| Time Series Discrepancies | |
| 2nd Highest Salary | |
| Empty Neighborhoods | |
| Rolling Bank Transactions | |
| Subscription Overlap | |
| Employee Salaries | |
| Merge Sorted Lists | |
| Comments Histogram | |
| Closest SAT Scores | |
| Top Three Salaries | |
| Monthly Customer Report | |
| Cumulative Distribution | |
| Experiment Validity | |
| 500 Cards | |
| Compute Deviation | |
| Find the Missing Number | |
| String Shift | |
| Maximum Profit | |
| Bagging vs Boosting | |
| Prime to N | |
| Session Difference | |
| Top 5 Turnover Risk |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report starting with a casual conversation about their resume, or with a hiring manager explaining the team's work and the role. One candidate said recruiters described a technical, hiring manager and HR sequence. Prepare a concise story of your background and ask what each later round will cover.
One candidate reported a HackerRank test with 30 questions in 45 minutes, described as straightforward but time-pressured. Another described a skills round on pandas, numpy, precision and recall, data types and data ethics. Which format you get may depend on the team, so practice both speed and concepts.
Candidates describe ML concept questions such as bias-variance, multicollinearity, tree-based models, boosting, F1, imbalanced data and cross-validation. One candidate expected coding but received a theory session, so confirm the format. Another reported an NLP and multiclass classification homework. Expect to explain your past projects and modeling choices in depth.
Candidates who reached a final stage report a behavioral conversation, in one case with about three managers, covering why you should be hired, customer interactions and how you organize work. One said it used the STAR format but felt relaxed. Prepare examples and project stories, and expect the tone to vary by team.
