← Back to Directory

DRW interview notes, page 2 of 2

Large Enterprises

Screna has 18 DRW interview write-ups. The most reported round is the technical phone screen (14).

18 interview notes · updated Jul 2026

DRW·Data Scientist·Technical Phone Screen

Jun 2026
DRW data scientist interview that was basically three algorithmic coding problems back to back. Pretty heavy on the CS fundamentals side for a DS role, which surprised me a bit.
  • Given an array of movie ratings, implement a solution that returns both the length of the longest strictly increasing subsequence and one valid subsequence as a list of 1-based indices. Must run in O(n log n) time.
  • Implement a data structure supporting two operations on an integer array: point updates (add a value to a single element) and range sum queries. Must handle up to 200,000 elements and operations efficiently.
  • You have a list of release timestamps and a list of deployment windows. For each window, assign the most recent available release that hasn't been used yet and is no later than the window's timestamp. If none exists, return -1. Aim for O((n+m) log(n+m)) time.

“This is the classic LIS problem but they wanted the actual indices too, not just the length, which tripped me up.”

View Post

DRW·Machine Learning Engineer·Technical Phone Screen

Jun 2026
DRW ML Engineer interview threw a serious quant-heavy coding problem at me, basically a full portfolio analytics library in pandas. The scope was way bigger than I expected for a single question.
  • Given one or more CSV files with daily asset prices or returns and optional portfolio weights, write Python and pandas code to load, clean, and align the data by date, handle missing values and outliers with documented choices, compute simple and log returns from prices, generate portfolio return time series with support for fixed-weight and periodic rebalancing, calculate cumulative return, annualized return, annualized volatility, Sharpe ratio, and maximum drawdown, support multiple portfolios with a summary table and exportable CSV and plot-ready output, and include unit tests for edge cases like non-overlapping dates, zero weights, and NaNs. Also report algorithmic complexity and justify your design decisions.

“This was a lot.”

View Post

DRW·Data Scientist·Technical Phone Screen

Jun 2026
Got a probability/geometry question at DRW for a Data Scientist role. Pretty clean problem but the setup trips you up if you're not careful about what you're actually computing.
  • Three points are chosen uniformly at random on the circumference of a unit circle. They divide the circle into three arcs. What is the expected length of the arc that contains the point (1,0)?

“My first instinct was to just say 2π/3 because three arcs, circle splits evenly on average, done.”

View Post

DRW·Machine Learning Engineer·Technical Phone Screen

Jun 2026
DRW ML Engineer interview with a quantitative coding problem that blends finance and Python. Pretty niche if you haven't touched portfolio optimization before, but manageable if you know your numpy.
  • You're given a DataFrame of price-return time series. Simulate random portfolio weights, compute expected return, volatility, and Sharpe ratio for each, then return the weight vector that maximizes the Sharpe ratio.

“I'd seen Monte Carlo portfolio stuff before but never had to code it under pressure.”

View Post

DRW·Machine Learning Engineer·Take-home Assignment

Jul 2026
DRW ML Engineer interview, looks like a take-home or practical coding round centered entirely on a constrained ML problem. No model swapping allowed, which forced me to actually think about the data side for once.
  • Given a fixed LinearSVC classifier, implement train() and test() functions so that your model beats a provided baseline accuracy on a hidden test set. You can only touch preprocessing, feature engineering, and hyperparameters.
  • How would you tune hyperparameters and validate your model without ever looking at the test set? Describe a strategy that lets you confidently claim you've beaten the baseline using only training data.
  • How do you handle class imbalance within the LinearSVC constraint? Walk through your approach.
  • How do you make your ML pipeline reproducible and trackable across experiments?

“The constraint is the whole puzzle.”

View Post

DRW·Machine Learning Engineer·Technical Phone Screen

Jul 2026
DRW ML Engineer interview was a deep conceptual grilling on Transformer internals. No coding, just pure theory for an hour straight. If you haven't touched the math behind attention in a while, this will humble you fast.
  • Derive the scaled dot-product attention formula and explain why the scaling factor is necessary.
  • Compare pre-LN and post-LN Transformer architectures and how each affects training stability.
  • When would you choose ReLU, GELU, or SiLU as an activation function, and how does each affect gradient flow?
  • Explain sinusoidal versus learned positional encodings and how each handles extrapolation to longer sequences.
  • What regularization techniques are commonly used in Transformers, like dropout and label smoothing, and when should each be applied?
  • Walk through your reasoning for optimizer and learning rate schedule choices, specifically AdamW with warmup and cosine decay.
  • Explain gradient clipping and the trade-offs involved in mixed-precision training.
  • What are the common causes of training divergence in deep learning, and how do you diagnose and fix them?
  • What is the difference between cross-attention and self-attention, and in what situations would you use one over the other?
  • How does attention masking work differently for causal language models versus bidirectional models?

“Knew the formula cold but fumbled explaining the WHY behind the scaling.”

View Post

DRW·Data Scientist·Technical Phone Screen

Jul 2026
Quant-heavy technical screen for a Data Scientist role at DRW. Three probability and stats problems back to back, the kind where you have to show actual derivations not just state answers. Felt more like a math exam than a conversation.
  • Given a finite-state, irreducible, aperiodic Markov chain with a provided transition matrix P, compute the stationary distribution π by solving πP = π subject to the probabilities summing to 1. Also explain why the solution is unique.
  • With X drawn from an Exponential distribution with rate λ and N drawn from a Poisson distribution with the same parameter λ, derive E[X] and E[N] from first principles, showing the integration and summation steps explicitly.
  • For a given 2x2 zero-sum game payoff matrix (payoffs to Player A), find the mixed-strategy Nash equilibrium and calculate the probability each player places on their first action.

“The mechanics of solving the linear system were fine but the uniqueness part tripped me up a bit.”

View Post

DRW·Data Scientist·Technical Phone Screen

Jul 2026
DRW data scientist interview with a linear algebra proof question that felt more like a math exam than anything I expected going in. Just the one question from what I can tell, but it had layers.
  • Given an invertible n×n matrix A with eigenvalues λ1 through λn, prove that the eigenvalues of A inverse are their reciprocals. Then find the sum of those eigenvalues and express it using a matrix operation on A inverse. What assumptions are required for this to hold?

“The proof part I actually got through okay.”

View Post