How Hudson River Trading Interviews in 2026: Process, Questions & What They Score

1. 01

   Count the positive integers below n (n up to 10^9) whose base-4 digits are all 0 or 1 (the 'fancy numbers' problem).

   Digit-based counting over the base-4 digits of n, since checking every number is too slow at 10^9.

2. 02

   Implement a game simulation such as 2048 or Reversi/Othello.

   Careful board-state simulation and rule edge cases, written correctly under a clock.

3. 03

   Implement a class that supports add(x) and median(), both with sub-O(n) complexity per call.

   Data structure design, such as two heaps or an order-statistic tree, for sub-linear add and median.

4. 04

   Write code that finds the best poker hand, live, in 45 minutes.

   Hand-ranking logic with counting and sorting of ranks and suits, coded live while narrating.

5. 05

   Filter sets of timestamped trades and calculate the resulting profits.

   Data processing: filtering timestamped records and aggregating profit correctly within the time limit.

6. 06

   Design a deque with O(1) push and pop at both ends and O(1) random access.

   Circular buffer or chunked array design that keeps end operations and indexing at O(1).

7. 07

   Given street lamps on a one-dimensional line, find how many lamps light each point, or the smallest integer point lit by the most lamps.

   Sweep line or difference array to count overlapping coverage intervals along a line.

8. 08

   How would you implement a hashmap, handle collisions, and keep amortized time complexity during resizing?

   Hash table internals: collision handling, load factor and the amortized cost of resizing.

9. 09

   Given N large sorted log files that do not fit in memory, merge them into a single sorted file.

   External k-way merge with a min-heap and buffered reads when the data exceeds memory.

10. 10

    Design a scheduler that triggers a callback after T milliseconds and supports cancelling it.

    Timer scheduling with a min-heap or timing wheel, plus bookkeeping for cancellation.

11. 11

    Implement a pool allocator in C++.

    Low-level C++ memory management: fixed-size blocks, a free list and alignment.

12. 12

    Wrap a stream reader that returns fixed-size chunks so that it supports reads of arbitrary length.

    Buffering and leftover-state handling to turn fixed-size chunk reads into arbitrary-length reads.

13. 13

    Find a local minimum in a one-dimensional array and then in a two-dimensional grid.

    Binary search on the local-minimum property, extended from one dimension to a grid.

14. 14

    Solve a tree path problem involving minimum excluded primes.

    Tree traversal combined with number theory, such as primes and minimum-excluded values on a path.

Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.

1. 01

   Explain the amortized complexity of std::vector push_back.

   Amortized analysis of dynamic array growth, capacity doubling and reallocation cost.

2. 02

   Explain threads versus processes and how virtual memory works.

   OS internals: isolation and sharing in threads versus processes, and paging in virtual memory.

3. 03

   What is the difference between stack and heap memory, and what is the memory layout of a process?

   Process memory layout and the lifetime and cost differences of stack versus heap allocation.

4. 04

   How does memory alignment affect struct sizes, and what is Small String Optimization?

   Struct padding and alignment rules, plus the small-buffer optimization in std::string.

5. 05

   What commonly causes segmentation faults in high-performance C++ code, and how do you handle them?

   Debugging C++ memory errors such as dangling pointers, out-of-bounds access and stack overflow.

6. 06

   Analyze the potential causes of latency spikes in a low-latency trading system.

   Systems diagnosis: cache misses, page faults, lock contention, allocation and scheduler jitter.

7. 07

   Two threads each increment a shared counter 1 million times, and the counter ends below 2 million. Why?

   Data races on non-atomic increments, and the fix with atomics or locks.

8. 08

   Design a service that consumes a real-time order flow feed, computes per-trader and per-strategy P&L and risk exposures, and alerts when limits are breached, within a 1 millisecond latency budget.

   Low-latency system design: event pipeline, in-memory aggregation state and a strict latency budget.

Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.

1. 01

   You are given an existing C++ client and server trading codebase that supports adding and cancelling orders. Implement order modification.

   Reading an unfamiliar C++ client and server codebase and extending it by adapting the add and cancel code.

2. 02

   Explore a provided dataset in a Python notebook using pandas and plotting (one intern report: a heart disease prediction dataset, documentation allowed).

   Exploratory data analysis with pandas and plots in a Jupyter notebook, reasoning about real data.

Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.

1. 01

   Romeo and Juliet each arrive at a uniformly random time within one hour. The first to arrive waits 15 minutes and then leaves. What is the probability they meet?

   Geometric probability on the unit square: the area where the two arrival times are within 15 minutes.

2. 02

   X and Y are independent and uniform on [0, 1]. What is the probability that X + Y is less than 1?

   Geometric probability: area under the line X + Y = 1 in the unit square.

3. 03

   A coin lands heads with probability p. What is the expected number of flips until you see two consecutive heads?

   Expected value by first-step recurrence over states: no heads yet, one head, two heads.

4. 04

   Solve a gambler's ruin or coupon collector problem.

   First-step recurrences for absorption probabilities and expected waiting times.

5. 05

   Answer a basic statistics question, then write code that runs a simulation to check your intuition.

   Analytic derivation followed by a Monte Carlo check that validates the result in code.

6. 06

   Derive the optimal strategy for a dice or card game, then write code that validates it by simulation.

   Expected-value optimization of a game strategy, then simulation to confirm it.

7. 07

   Propose an estimator for a given quantity and say whether it is biased.

   Computing an estimator's expected value to test bias, as with the maximum of uniform samples.

Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.

1. 01

   Tell me about a time you disagreed with a technical direction. How did you challenge it, and how did you resolve it?

2. 02

   Walk me through the most complex system you have built, including the trade-offs you made.

3. 03

   Walk me through your background and why you want to work at HRT.