Capital One Interview Questions — Page 25

Capital One·Data Scientist·Technical Phone Screen

Jun 2026

Capital One Data Scientist interview with a pretty gnarly SQL question that combined data cleaning, window functions, and multi-part output into one big query. No behavioral stuff from what I can tell, just pure technical depth.

You're given four CSVs (platforms, ads, videos, and a daily totals table) where the date column uses dashes to mean 'same as the row above.' Using only SQL (no procedural code), forward-fill those dates, then for a 7-day window compute per-ad and per-platform aggregates including CTR, and return both the top 3 ads by plays and the peak watch-time date per ad, all in a single query.

“This thing had like five sub-problems stapled together and they wanted one query.”

View Post

Capital One·Machine Learning Engineer·Technical Phone Screen

Jun 2026

Capital One MLE round that was heavier on product reasoning and rule-based logic than I expected. No ML at all, which threw me off a bit. The virtual credit card case was interesting but the bit-encoding in card numbers felt like it came out of nowhere.

What are the key benefits and drawbacks of virtual credit cards, from both the end user's perspective and the bank's perspective?
Given a transaction with encoded attributes in the card number and transaction ID digits, describe a data model and validation flow to determine whether the transaction is valid under Visa or Master network rules.
Using card_number '1234567891011111' and transaction_id '50781100' with amount 150, is the transaction valid?
How would you write unit tests and design edge cases to validate this transaction logic?

“I actually liked this one.”

View Post

Capital One·Data Scientist·Technical Phone Screen

Jun 2026

Capital One data scientist interview, technical screen focused entirely on a single meaty SQL problem with sample data. The question looked straightforward at first glance but the returns logic and tie-breaking rules added enough wrinkles to slow me down.

Given a fixed 7-day window (2025-08-26 to 2025-09-01), write a single SQL query that finds the top product by net revenue within each category. Net units must subtract returns that fall in the same window, and ties should be broken by net units then lexicographic product ID. Use window functions to pick one winner per category. Output category, product_id, net_units, and net_revenue_usd.

“The returns join is where I fumbled.”

View Post

Capital One·Data Scientist·Technical Phone Screen

Jun 2026

Capital One data scientist interview that went pretty heavy on the quantitative side. The case question was a full profitability build for a hypothetical vegan burger launch, which felt more like a finance interview than anything I expected from a DS role.

You're evaluating whether to launch a vegan burger line. Given fixed training costs, a monthly supplier retainer, per-unit variable costs, a selling price, and a projected sales volume of 120 million units in Year 1, compute the incremental operating profit for the year. Make sure to annualize the monthly retainer.
Break down that operating profit into its fixed and variable components. What is the per-unit contribution margin, the contribution margin ratio, and the total contribution?
What is the breakeven unit volume for this vegan burger line, and what does that number mean in practical terms?
Run a sensitivity analysis: if actual sales volume comes in 10% above or below the 120 million unit forecast, what does incremental profit look like in each scenario?

“I fumbled the retainer at first.”

View Post

Capital One·Data Scientist·Technical Phone Screen

Jun 2026

Capital One DS interview that was basically a finance case disguised as a product question. Four parts, all building on each other, and the math gets uncomfortable fast if you lose track of what's actually being discounted.

Given a base of 2 million cardholders where 5% use the Rent-a-Home product annually at $500 average spend, and a proposed 30% discount fully funded by Capital One, write the break-even equation where incremental revenue from new cardholders equals the total promo cost across existing and new RH users. Solve for the number of new customers needed.
Express the break-even new customer count as a percentage of the 2 million cardholder base, and assess whether that acquisition target is realistic given typical funnel conversion and credit approval rates.
Extend the model so that existing RH users increase spend by 20% during the promo (to $600) and new customers acquired through the promo spend $1,000 on RH. With the 30% discount still fully funded by Capital One and first-year net revenue per new cardholder still at $550, solve the updated break-even equation: 550x minus 0.30 times (1000x plus 2,000,000 times 5% times 600) equals zero.
Compare the break-even new customer count from the extended scenario to the original, and explain intuitively why the required number shifts even though spend is higher.

“The setup looks clean until you realize you have to account for promo cost on both existing users and the new ones simultaneously.”

View Post

Capital One·Software Engineer·Onsite - Behavioral / Leadership

Jun 2026

Behavioral round at Capital One for a software engineer role. Pretty much one meaty question with a bunch of follow-ups hanging off it, so if you don't have a solid story ready you'll run out of things to say fast.

Tell me about a time you had to learn something new quickly under pressure. How did you approach it, what resources did you use, how did you apply what you learned, and what was the result?

“This is the kind of question that sounds easy until you're actually in it and realize your story has holes.”

View Post

Capital One·Software Engineer·Onsite - Behavioral / Leadership

Jun 2026

Capital One software engineer behavioral round. Three main themes: learning fast, pushing back on decisions, and handling real pressure. Pretty standard for a big bank but the follow-up questions they prep you for are no joke.

Tell me about a time you had to pick up something new quickly. How did you go about it, and what happened?
Describe a time you pushed back on an existing process or decision. How did you bring others along, and what actually changed?
Walk me through a significant challenge you faced at work. What did you do, what did you give up to get there, and what did you take away from it?

“This one sounds easy until you realize they want the follow-ups too: scope, what you cut corners on, whether it actually worked.”

View Post

Capital One·Data Scientist·Take-home Assignment

Jun 2026

Capital One Data Scientist take-home where they basically handed you four CSVs and said 'replicate what an analyst would do in Excel, but in pandas.' The scope was bigger than I expected for a single assignment.

Load four CSV files into pandas in a way that handles a file path failure gracefully, without crashing the entire script.
Compute net revenue per order by subtracting any refund amount from the order total, treating orders with no refund record as having a zero refund.
Aggregate net revenue by month and acquisition channel for a specific three-month window, using the order date as the basis for month assignment.
Left-join the aggregated monthly channel revenue to a targets table and compute the gap between actual net revenue and the revenue target, treating channels or months with no target as a zero target.
Produce two pivot tables: one showing net revenue by month and channel, and another showing the revenue target gap by month and channel. Channels with no orders in a given month should still appear with a value of zero.

“I went with a try/except loop and a fallback path, which felt right, but I spent way too long debating whether to use a retry decorator or just catch the exception inline.”

View Post

Capital One·Machine Learning Engineer·Onsite - Behavioral / Leadership

Jun 2026

Interviewed for an ML engineer role at Capital One, one behavioral question about handling pressure came up and that was pretty much the whole thing.

How do you manage projects when you're under pressure?

“Pretty standard behavioral question but I still fumbled it a bit.”

View Post

Capital One·Data Scientist·Technical Phone Screen

Jun 2026

Capital One Data Scientist interview that leaned heavier on business math than I expected. The break-even stuff felt more like a finance case than a data science problem, but I guess they want to know you can think about unit economics.

A streaming company has monthly fixed costs of $100M, a variable cost of $4 per subscriber per month, and charges $9 per subscriber per month. Calculate the monthly break-even subscriber count, showing your formula and units clearly.
If the price per subscriber increases to $10 while all costs stay the same, what is the new break-even subscriber count?
If fixed costs rise to $120M while price stays at $9 and variable cost stays at $4, recalculate the break-even count. Also state any assumptions you're making about cost timing and recognition.

“Straightforward once you remember the formula: fixed cost divided by (price minus variable cost).”

View Post