Citi Data Scientist candidates report an HR call, technical rounds on statistics, machine learning and resume projects, and in one case a four-interview superday. Coding is described as basic.
$125K
Avg. Base Comp
$141K
Avg. Total Comp
Not reported
Typical Rounds
Not reported
Process Length
Candidates describe the Citi Data Scientist interview as fundamentals-heavy and resume-driven. Across three Data Scientist reports, interviewers asked how well you understand a concept, not just whether you can name it.
One candidate was asked about linear regression, ROC, AUC, specificity and F1, and how Random Forests work. They were also asked about comfort with Python, R and SAS. Another reported autocorrelation, ACF and PACF plots, Monte Carlo simulation, decision trees, random forests and confusion matrices. That candidate also saw RAG systems, vector databases, SQL and Python OOP. A third described one SQL question, one pandas question (finding the song that ranked first most often), a root-cause-analysis exercise, and a close walk-through of past projects.
Project discussion came up in two of these reports. One candidate said interviewers cared more about the problem solved, the method and the findings than the tech stack. Prepare two or three projects you can explain end to end, including why you chose your approach and what you would change.
Coding looks like a smaller share. One candidate called the algorithm questions straightforward, mentioning linked lists and basic data structures.
Structure varied. One candidate had HR, then two teammates, then a manager, with an offer inside a month. Another went through 2-3 technical rounds, an HR round and a four-interview superday. Two of the three Data Scientist candidates in this set reported compensation conversations late in the process, so check your expectations early. All three reported offers.
Synthesized from 5 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 Citi process.
I went through a multi-round technical interview process at Citi for a Data Scientist role. The process involved 2–3 technical rounds followed by an HR round, and I ultimately moved into a superday format consisting of 4 back-to-back interviews with team members and a department director.
The technical rounds heavily emphasized statistics and machine learning fundamentals. I was asked to explain concepts like autocorrelation, ACF and PACF plots, and Monte Carlo simulation. Questions also covered classic ML algorithms—decision trees, random forests, and confusion matrices featured prominently. I had to discuss projects from my resume in detail, explaining my approach and thinking. There was also a focus on RAG (Retrieval-Augmented Generation) systems, including vector databases and LLM integration, which felt like a newer addition to their stack. SQL and Python OOP came up as well, testing both database and language proficiency.
Algorithmic challenges were present but generally straightforward—I saw references to linked list construction and basic DSA problems, nothing that demanded advanced algorithm optimization. The interviews felt conversational rather than adversarial, especially in the earlier rounds, though the superday format with multiple senior interviewers raised the intensity.
One thing that stood out: salary discussion happened late in the process. I made it through the entire process only to be presented with a significantly lowball offer afterward, which was frustrating. I'd recommend clarifying compensation expectations upfront before committing to multiple rounds. The technical bar seemed reasonable for a mid-level DS role, but my overall experience left me wary given how compensation was handled at the end.
Prep tip from this candidate
Focus on statistical concepts—especially ACF/PACF plots, time series interpretation, and Monte Carlo simulation—as these came up repeatedly. Also prepare concrete examples of ML projects from your resume, not just the names. RAG systems and vector databases appeared in recent cycles, so familiarize yourself with that architecture. Clarify salary expectations and total comp before accepting the interview loop.
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 Citi
Write a query to show the number of users, transactions, and total order amount per month in 2020
| Question | |
|---|---|
| Attribution Rules | |
| Size of Joins | |
| Jars and Coins | |
| Above Average Product Prices | |
| New Partner Card | |
| International e-Commerce Warehouse | |
| Replace Words with Stems | |
| Dijkstra implementation | |
| Hurdles In Data Projects | |
| Logistic Regression from Scratch | |
| Concurrent LLM Serving | |
| Digital Marketing Metrics | |
| Loan Model | |
| Client Solution Pushback | |
| Offer Matching API Design | |
| String Palindromes | |
| Index Fund Return | |
| Text Editor With OOP | |
| Seller Type Modeling | |
| Why Do You Want to Work With Us | |
| Your Strengths and Weaknesses | |
| Company Acquisition Choice | |
| Concentric Circles | |
| Data Cleaning Experiences | |
| Singly Linked List | |
| Student Tests | |
| Measuring Text Difficulty | |
| Risk Model for a Mortgage Bank | |
| Optimizing Threshold Adjustment in Default Risk Models |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report starting with an HR call or HR screening round before any technical interview. One candidate also described a couple of HR touchpoints between technical rounds to align on logistics and expectations. Prepare a concise background summary and be ready to discuss role fit.
One candidate interviewed with two future teammates and was asked about linear regression, data manipulation, ROC, AUC, specificity and F1. Another report mentions autocorrelation, ACF and PACF plots, Monte Carlo simulation, and tree-based models. Expect questions on how deeply you understand each concept.
Candidates report a mix of practical coding: one SQL question and one pandas question in one case, and Python OOP, SQL and basic data structures such as linked lists in another. One candidate described the algorithm questions as straightforward. Python, R and SAS comfort also came up once.
Candidates report detailed resume questions, including a root-cause-analysis question and close questioning on past projects. One candidate said interviewers cared about the problem solved, how it was solved and the findings more than the tools. Another had a second technical round with a manager at a similar difficulty.
One candidate reported a superday of four back-to-back interviews with team members and a department director. Others described a salary conversation near the end. Process length varies by report, and one candidate received an offer within a month. Clarify compensation expectations early.
