Upstart Interview Questions — Page 2

Upstart·Software Engineer·Technical Phone Screen

Apr 2026

Interviewed for a software engineering role at Upstart and got a data manipulation problem involving stock prices. Pretty straightforward on the surface but took a minute to think through the aggregation logic cleanly.

Given two lists, one with company names and one with daily stock prices for each company, find the top 3 companies by average stock price.

“My first instinct was to zip the two lists together and compute averages, which worked fine.”

View Post

Upstart·Data Scientist·Technical Phone Screen

Apr 2026

Upstart data scientist interview had a probability/statistics question that looks deceptively simple but is actually a sampling bias trap. The math itself isn't hard once you see what's going on, but I almost blew it by treating the school survey as a direct household estimate.

You survey 100 children at a school and ask how many kids are in their family. 50 say 1, 20 say 2, 30 say 3. Then you knock on a random house in the same town and ask how many children live there. What's the probability it's exactly 1 child?
Why doesn't the school survey directly tell you the household distribution? What's the statistical issue?
How would your answer change if households with zero children existed, or if families larger than 3 were possible but just not observed in the sample?

“My first instinct was 50% because 50 out of 100 kids said they were in a 1-child family.”

View Post

Upstart·Data Scientist·Technical Phone Screen

Apr 2026

Got a technical screen for a Data Scientist role at Upstart that went deep into logistic regression fundamentals. One meaty multi-part question covering the full ML loss function pipeline, from likelihood to gradients. Felt like a written exam more than a conversation.

For binary logistic regression, write out the maximum-likelihood objective both with and without L2 regularization, and be explicit about whether you're maximizing or minimizing.
Write out the explicit cross-entropy loss for a binary logistic regression model.
Derive the gradients of the cross-entropy loss with respect to the weight vector and the bias term.
What is the difference between the 'loss', the 'objective', and the 'regularized objective' in this logistic regression setup?

“I defaulted to writing the minimization form (negative log-likelihood) without being asked, which is fine but I fumbled the transition when they asked me to show the maximization version too.”

View Post

Upstart·Data Scientist·Technical Phone Screen

Apr 2026

Upstart data science interview with a probability/statistics brain teaser that looks straightforward until you actually think about the sampling mechanism. The whole thing hinges on whether you catch the bias in how the kids were selected.

You survey 100 kids at a school and ask how many children are in their family. 50 say 1, 20 say 2, 30 say 3. You then knock on a random house in town and ask how many kids live there. What's your best estimate of the probability that house has exactly 1 child? State any assumptions you need about how the kids were sampled.

“This one got me.”

View Post

Upstart·Data Scientist·Technical Phone Screen

Apr 2026

Upstart data science interview with a meaty credit risk scenario that pushed into territory I hadn't fully prepped for. The question sounds clean on the surface but there's a lot of depth hiding underneath it.

Your current credit risk model was trained exclusively on applicants with credit scores of 650 or above. The business now wants to extend lending to people below that threshold. How would you use the existing model and historical data to score this new population?

“I started talking about retraining and the interviewer kind of let me run with it before gently steering toward the covariate shift angle, which I'd glossed over.”

View Post

Upstart·Data Scientist·Technical Phone Screen

Apr 2026

Upstart data scientist interview, looked like a technical phone screen focused almost entirely on probability, stats, and regression fundamentals. Nothing about the product or business, just math. Felt more like a grad school exam than a job interview.

A radioactive atom has a half-life of 1 day. Starting with 100 identical atoms, what is the probability a single atom survives after 10 days, what distribution describes the number of surviving atoms, and what are the expected number remaining and the probability at least one is left?
Given the prior probability of a condition, the probability of a positive test result given the condition is present, and the probability of a positive result given the condition is absent, derive the posterior probability that the condition is present given a positive test.
If y = x + e where x and e are independent standard normals, what is the slope from regressing y on x, and what is the slope from regressing x on y?
In the Monty Hall problem, after the host opens a losing door, should you switch or stay, and what is the winning probability for each strategy?
You roll a fair n-sided die repeatedly. What is the expected number of rolls needed to see every face at least once?
What is a likelihood function in parametric modeling, and how does it differ from a probability statement?

“The single-atom survival probability is just (1/2)^10, which is clean enough.”

View Post

Upstart·Machine Learning Engineer·Onsite - Coding / Algorithms

Apr 2026

Upstart ML engineer interview with a hands-on case study using the California housing dataset. The whole thing was less about throwing features at a model and more about showing you understand distribution shift and can actually diagnose what's going wrong.

You're given the California housing dataset where train and test sets come from different income brackets, causing feature and label distribution shift. Build a regression model from scratch to predict house prices, then analyze and optimize it with that shift in mind.

“I spent way too long on feature engineering early on and kind of assumed the model analysis part would be quick.”

View Post

Upstart·Software Engineer·Technical Phone Screen

Apr 2026

Coding round for a Research Scientist role at Upstart. One question, pretty focused on simulation and probability, the kind of thing where you need to actually write working code and also know the math behind it.

Write a simulation of a particle decay process: start with 100 particles, each day every surviving particle independently decays with probability 1/2, run for 10 days across many trials (around 100,000), and estimate the probability that at least one particle survives. Then compare your Monte Carlo result to the exact analytical answer.

“The coding part was fine, nested loops, count survivors, average across trials.”

View Post

Upstart·Data Scientist·Onsite - Multi Round

Apr 2026

Pretty intense interview for a Data Scientist role at Upstart. Three big questions that each could've been their own 45-minute conversation, and the interviewer clearly wanted structured thinking with real numbers attached, not hand-wavy answers.

Walk me through your career history in order. For each job change, what was pulling you toward the new role and pushing you away from the old one? What did you expect to get out of the move, and did the numbers back that up?
Using a specific competitive analysis framework, explain what gives your current company a durable advantage in the market. Back it up with metrics, customer stories, or benchmarks, and identify one real threat to that advantage and how the company is addressing it.
Tell me about a time you had to decide how much effort to put into a high-stakes presentation under a tight deadline. What scope trade-offs did you make, did you push back on anyone, and what actually changed as a result of that presentation?

“This is the kind of question that sounds easy until you're actually doing it.”

View Post

Upstart·Software Engineer·Technical Phone Screen

Apr 2026

Coding screen at Upstart for a software engineer role. One filtering and aggregation problem, nothing too exotic, but the edge cases are where it gets you.

Given a list of name-score pairs, filter out entries where the score exceeds a threshold, then find and return the name with the highest score among the remaining entries.

“Seemed straightforward at first.”

View Post