Capital One Interview Questions — Page 7

Capital One·Software Engineer·Onsite - System Design / Architecture

May 2026

Capital One SWE interview that was basically one big system design problem: build a banking system from scratch with account creation, deposits, transfers, and some kind of leaderboard for most active accounts. Input came through CSV or JSON parsing which added a layer I didn't fully expect.

Design a banking system that supports account creation, deposits, transfers between accounts, and a way to retrieve the most active accounts ranked by total monetary activity. Commands are issued by parsing structured input like CSV or JSON.

“The parsing requirement threw me a bit.”

View Post

Capital One·Software Engineer·Technical Phone Screen

May 2026

Capital One technical screen, one coding problem that looked simple on the surface but had enough edge cases to trip you up if you weren't careful. Nothing crazy, just needed to think it through properly.

Given a string, if both the first and last characters are vowels (case-insensitive), reverse the substring between them (excluding the endpoints) and return the result. Otherwise return the string as-is. Expected O(n) time with a two-pointer approach, and you need to walk through edge cases like empty strings, single characters, two characters, and mixed case.

“The reversal itself is not the hard part.”

View Post

Capital One·Data Scientist·Take-home Assignment

May 2026

Capital One data scientist take-home that was basically a full pipeline design problem. Five parts, covers everything from SQL feature engineering to orchestration and serving. More like a mini project than a typical interview question.

Given raw clickstream events and a users table, write SQL or Python to produce user-day features including session counts (using a 30-minute inactivity gap), page views, cart additions, purchases, revenue, days since signup, country, and top traffic source. Make sure to deduplicate by event_id and correctly bucket timestamps to UTC dates.
How do you guarantee idempotency and handle late-arriving or out-of-order events in this pipeline? Walk through your approach to things like watermarking, upsert/merge strategies, and the tradeoff between partition overwrite and append-only designs.
Describe your backfill plan for a roughly three-month date range. How would you re-run only the affected partitions safely without corrupting existing data?
What concrete data quality checks would you add to this pipeline, and write two unit tests that would catch common bugs in this kind of feature engineering code.
Outline the orchestration DAG for this pipeline, including task dependencies. Also describe your storage format and partitioning choices, and explain how you'd expose the output for both model training and online inference.

“This was the meatiest part.”

View Post

Capital One·Data Scientist·Technical Phone Screen

May 2026

Capital One data scientist technical screen, all four questions were SQL plus pandas combos on the same schema. Pretty intense for a phone screen, felt more like a take-home crammed into a live session.

Given a customers table with missing age values, impute each null using the median age within that customer's tier. If a tier has no non-null ages to compute a median from, fall back to the global median. Return customer_id and the imputed age.
You have a staging_events table that may contain new rows or updated versions of existing rows relative to a live events table. Write an upsert: insert rows with event_ids not already in events, and for duplicate event_ids with differing values, keep only the row with the latest event_date. Return the final deduplicated events table.
For the 7-day window ending today, compute net revenue per tier. Purchases count positive, refunds negative, and non-monetary events like page views should be excluded. Use the imputed ages from the first question and restrict to customers aged 18 to 65. Return tier, total revenue for the window, and distinct customer count.
Among customers who had at least one monetary event in the last 7 days, what fraction also had any event in the prior 7-day window? Return one row per tier with the retention rate. Also describe what indexes you'd add and how you'd test correctness.

“This one tripped me up more than it should have.”

View Post

Capital One·Data Scientist·Technical Phone Screen

May 2026

Capital One DS interview had a role-play exercise where I had to explain model degradation post-deployment to a non-technical PM, which sounds manageable until you're actually doing it live and second-guessing every word choice.

You're a Data Scientist and I'm a Product Manager. Our model performs well on the validation set but noticeably worse after deployment. Walk me through what might be causing that and what you'd do about it.

“This one tripped me up more than I expected.”

View Post

Capital One·Data Scientist·Technical Phone Screen

May 2026

Capital One Data Scientist case interview built around a restaurant-and-coupon-site scenario. The whole thing was a structured profitability case with both qualitative and quantitative components, which I wasn't fully expecting for a DS role. Felt more like a consulting case than anything analytical.

Before deciding to partner with a discount voucher site, what business and financial factors would you consider?
What quantitative levers would you model when evaluating a coupon partnership?
Calculate the restaurant's current daily profit with no partnership in place.
A voucher sells for $15 and gives $30 of in-restaurant credit. The site takes a 40% commission on the voucher sale. What minimum average spend per voucher table breaks even on a marginal basis? Also compute the break-even if the 40% commission applies to the full check instead.
Based on the break-even spend you calculated, would you join the coupon site? Why or why not?
After joining the site you see 25 tables per day, 10 using vouchers, average spend at $36, same cost structure. Compute the new total daily profit and decide whether to continue the partnership.
Why can profit go down even when both the number of tables and average spend per table increased?
If the partnership must continue, what tactics would you suggest to improve profitability?

“I went straight to price elasticity and incrementality vs.”

View Post

Capital One·Data Scientist·Online Assessment (OA)

May 2026

Capital One online assessment for a Data Scientist role, two coding questions. Both were algorithmic, which I wasn't fully expecting for a DS position, but here we are.

Given a string, you can optionally reverse either a prefix or a suffix of any length (at most one operation total). Return the lexicographically smallest string possible.
Simulate a multi-lap race where after each lap you eliminate the driver with the highest cumulative time (tie-break alphabetically). Return the elimination order.

“My first instinct was to brute force every possible prefix and suffix reversal and just take the min.”

View Post

Capital One·Data Scientist·Technical Phone Screen

May 2026

Capital One data scientist interview that was basically one big pandas engineering problem. No behavioral fluff, just a dense denormalization task with a bunch of edge cases baked in. Left feeling okay about the joins but less sure about the payment_method tie-breaking logic.

Using only pandas (no row-level loops), write a function that merges seven input tables into a single denormalized fact table with a specific column order, handling aggregations like summed refunds per item, latest paid timestamp per order, earliest shipment per order, and a tie-breaking rule for payment method selection.
Write unit-test style assertions to verify that the output preserves row count from order_items, has no duplicate order_item_id values, and returns columns in the exact specified order.

“The joins themselves weren't the hard part.”

View Post

Capital One·Data Scientist·Technical Phone Screen

May 2026

Capital One data scientist interview built around a flight delay prediction scenario. The whole thing was one big multi-part case that went from data cleaning all the way to experimental design, which I wasn't fully expecting.

You're given a raw dataset for a flight delay prediction model. What data quality issues would you look for and how would you address them?
Should this flight delay problem be framed as a classification task or a regression task, and why?
VIF scores indicate high multicollinearity in your features. How would you diagnose and explain this to a fellow data scientist, and what steps would you take to fix it?
How would you design an experiment to help resolve or confirm the multicollinearity problem you identified?

“I rattled off missing values and called it a day at first, then they kind of waited and I realized they wanted more.”

View Post

Capital One·Data Scientist·Technical Phone Screen

May 2026

Capital One data scientist interview with a deep-dive experimentation question around a credit card acquisition redesign. Single question but it covered basically every angle of A/B testing you can think of, from randomization to fair lending compliance.

You're launching a new credit card acquisition flow with updated APR disclosure and a signup bonus. Walk through a complete A/B test design: randomization unit, eligibility exclusions, primary and guardrail metrics, minimum detectable effect, statistical power, sample size, and expected duration. Also address selection bias from pre-approval and underwriting, cross-channel interference, peeking, regulatory constraints like fair lending, and how you'd analyze heterogeneous treatment effects by segment while controlling the false discovery rate.

“This question is basically a final exam for experimentation.”

View Post