Robinhood Data Scientist Interview Questions

Robinhood·Data Scientist·Technical Phone Screen

Jul 2026

Robinhood data scientist interview with a coding problem built around streaming tick data and OHLC aggregation. The question was more engineering-flavored than I expected for a DS role, which threw me off a bit.

Write a Python function that parses a string of price:timestamp pairs, buckets them into 10-second intervals, computes OHLC (open, high, low, close) for each bucket, and forward-fills any empty intervals using the last known price.

“I didn't expect a pure coding question this close to systems/data engineering territory for a DS interview.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

Jul 2026

Technical screen for a Data Scientist role at Robinhood, focused on interval-based algorithm problems tied to fraud detection use cases. The problem felt more like a software engineering question than a data science one, which threw me off a bit.

Given a list of user login sessions as [start_time, end_time] pairs, first determine whether any sessions overlap, then return the minimal set of session IDs to remove so the remaining sessions are non-overlapping and maximize total time covered.

“I recognized the overlap-detection part pretty fast, sort and scan, easy.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

Jun 2026

Robinhood fraud/risk analytics round for a Data Scientist role. Three SQL questions, all revolving around a transactions table, and they got progressively nastier toward the end. Not a vibe-check round at all, they wanted you to actually write the queries.

Write a SQL query to find the top 3 users with the highest total declined transaction amount within the last 7 days.
For each user and day, flag whether they had more than two transactions where the amount is NULL or greater than 500 within any 24-hour rolling window. Return user_id, window_start, and a risky_flag column.
Add a risk_level column to the transactions table using conditional logic: label amounts above 800 as high, amounts between 500 and 800 as medium, and everything else as low. Return 10 sample rows ordered by most recent timestamp.

“Pretty approachable as a warmup.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

Jun 2026

Robinhood data science interview that went deep into causal inference for a trading metric decline. Four questions, all technical, no fluff. The kind of loop where you leave wondering if you actually knew what you were doing or just sounded like you did.

A new product release went out on July 10, 2025. Executed trades per active user has dropped since then. How do you design a difference-in-differences study to figure out if the product caused it, not just market conditions? Walk through your treatment and control groups, model spec, fixed effects, and how you'd test the parallel trends assumption. If pre-trends fail, what do you do?
Use at least two structural break detection methods, like CUSUM or Bai-Perron and Bayesian Structural Time Series, to quantify the effect size of the drop. What do you do when the two methods give you different answers?
Calculate the minimum detectable effect for a 10% drop in executed trades per active user. Assume daily data, alpha of 0.05, power of 0.80, 200,000 mean active users per day, baseline mean of 1.0 trades, and standard deviation of 1.5 trades. Show the pooled-variance t-test formula and state how many days of data you'd need.
What robustness checks would you run on this analysis? Think about placebo dates, randomization inference at the symbol level, wild bootstrap standard errors, and sensitivity to volatility controls like VIX or SPX returns and holiday effects. Also define clear pass/fail criteria that would determine whether to ship a fix.

“This is where I spent most of my mental energy.”

View Post

Robinhood·Data Scientist·Onsite - System Design / Architecture

Jun 2026

Robinhood data scientist interview that went deep into streaming infrastructure. The whole thing was basically one long system design question dressed up in data engineering clothes, which I wasn't fully expecting for a DS role.

Walk through a real-time analytics pipeline that ingests website click events via Kafka, processes them in Flink, and writes aggregated results to a data warehouse. Cover topic partitioning, windowing, state management, downstream table modeling, failure handling, and back-pressure strategies.

“This was a lot to unpack and I think I handled the first half better than the second.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

May 2026

Robinhood data scientist interview with a product analytics case focused on diagnosing a metric drop. Pretty standard fintech case but the layered follow-ups made it trickier than expected.

The share of new users who fund their account on day one has declined. Walk through how you'd systematically diagnose this, including what external and internal factors you'd look at, and what you'd explore if those all check out.

“I started by splitting the metric into numerator and denominator, which felt right, but then I kind of rambled through possible causes without a real structure.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

May 2026

Robinhood Data Scientist interview with a SQL question centered on membership period filtering. Pretty straightforward if you're comfortable with date range joins, but the 'Gold' tier filter is a small wrinkle that could trip you up if you're moving fast.

Given a customer profile table with membership type and date range, and a transactions table, write a SQL query that returns all transactions made during a customer's active Gold membership period.

“The join on customer_id is obvious, but I almost forgot to filter membership_type to 'Gold' before worrying about the date range.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

May 2026

Robinhood data scientist interview with a SQL and Python focus, all centered on a payments transaction schema. Nothing too wild but the window function follow-up was a step up from what I expected in terms of complexity.

Using a users table and a transactions table, write a SQL query that joins them and returns each user's name alongside their total amount sent and total amount received.
Extend your previous SQL query to include a running cumulative sum of amount_sent per sender, ordered by transaction_id, using a window function.
Given a transactions dataframe in Python, produce two Series showing the top 5 user IDs by number of transactions sent and the top 5 by number of transactions received, accounting for the fact that negative amounts reverse the direction of money flow.

“Pretty standard aggregation question.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

May 2026

Robinhood data scientist interview with a combined SQL and Python problem built around a transactions dataset. The SQL part had two tasks back to back and the Python part was more of an algorithm efficiency question dressed up as analytics. Pretty dense for a single session.

Given a Users table and a Transactions table, write a single SQL query that returns each transaction with normalized sender and receiver IDs, the normalized amount, and the country of each party. Transactions with a negative amount have their direction reversed, so you need to handle that with a CASE expression.
Using window functions on the normalized transaction data, produce a result showing each user's sent transaction count, their dense rank within their country by that count, and then filter to only the top 3 users per country.
Implement a Python function that takes an iterable of transaction records (each with an ID, a from:to user pair string, and an amount), applies the same direction-reversal rule for negative amounts, and returns the top 5 senders and top 5 receivers by transaction count. The solution should run in O(n) time and use O(u) extra space where u is the number of unique users.

“The direction-reversal rule tripped me up at first.”

View Post

Robinhood·Data Scientist·Technical Phone Screen

May 2026

Robinhood data scientist interview focused on product health metrics after a feature launch. Three connected questions that built on each other, which I didn't fully anticipate going in.

After launching a new product feature, what core metrics would you track to evaluate its health, and how would you compute them?
A key metric drops 5% suddenly. Walk through the possible root causes and explain what data you'd pull to validate each one.
Design an experiment to test a UI change meant to improve that metric. Cover the unit of randomization, your hypothesis, how long you'd run it, and what success looks like.

“I went straight to DAU and conversion rate, which was fine, but I fumbled explaining how to actually compute them in context.”

View Post