Hudson River Trading opens most technical hiring with a timed coding test, and recent candidates for software engineering internships describe it as three or four hard algorithm problems, often on CodeSignal and sometimes on HackerRank, in roughly 70 to 90 minutes. Algo developer candidates report a longer or harder variant. HRT job postings say AI tools are strictly prohibited during interviews and assessments unless you are told otherwise.
HRT describes its process in general terms, so this guide separates what HRT has published from what candidates report, as checked on November 16, 2026, and gives a 14-day plan for a test that rewards speed and exactness.
What HRT officially says
HRT's tech blog post on preparing for its software engineer interview, published September 15, 2021, describes the campus process as:
- A take-home coding test, usually timed and with a deadline, typically run on HackerRank or Codility. Language choice depends on the role, and some roles restrict it.
- About two phone interviews: roughly 45 minutes on systems knowledge, data structures or problem solving, and a programming interview in C++, Python or a language you know, depending on the role.
- A full day of back-to-back interviews, virtual or in person, on programming, systems knowledge such as memory, I/O and process management, problem solving, and how you collaborate and communicate.
The post adds that HRT revises the process regularly and that your recruiter is the main source on each round. In 2021 it said books and the internet were fine as a language reference during the coding test. That predates today's AI tools, and HRT's more recent job postings, as reproduced on job boards in 2026, say the use of AI tools during interviews or assessments is strictly prohibited unless otherwise instructed or agreed, that HRT uses various methods to check that responses are authentic, and in some postings that AI assistance found at any stage can mean disqualification or a rescinded offer. Follow the rules in your own invitation, and do not assume the 2021 allowance still applies.
The HRT online assessment, as candidates report it
| Element | What candidates report |
|---|---|
| Platform | CodeSignal in recent software engineering intern reports; HackerRank in others. |
| Software engineering | Four questions in about 70 minutes is the clearest recent CodeSignal report; other reports describe three or four problems on HackerRank in about 70 to 90 minutes. |
| Scoring shown | A Summer 2027 intern candidate, reporting in August 2026, describes a score out of 600. One prep site warns that the maximum score should not be mistaken for a cutoff. |
| Algo developer | Longer or harder: one guide describes about 150 minutes for three problems, another two or three hard problems in 90 minutes. |
| Difficulty | Medium to hard, sometimes competition style, with a ramp from a warm-up problem to hard ones. |
| Topics | Graphs, dynamic programming, strings and number theory, with weight on correctness and speed. |
| Language | Python or C++ in most reports; one quant analyst variant is reported as C++ only. |
These are self-reported accounts and prep-site summaries that differ by track and year. The four-question, roughly 70-minute CodeSignal format is the most common recent report, not an official specification. Our CodeSignal assessment guide explains how its scoring and proctoring work.
Topics, with practice examples
HRT does not publish its questions, and none of these are claimed to be on the test. They are well-known LeetCode problems pitched at the reported level, including one with an order-book flavor.
| Topic | Practice examples | What it trains |
|---|---|---|
| Graphs | Network Delay Time, Cheapest Flights Within K Stops, Swim in Rising Water | Dijkstra, bounded BFS and binary search on the answer |
| Dynamic programming | Edit Distance, Longest Increasing Subsequence, Burst Balloons | Recurrences over pairs and intervals |
| Strings | Minimum Window Substring, Longest Palindromic Substring | Windows and expansion |
| Number theory | Count Primes, Pow(x, n), Super Pow | Sieves and modular arithmetic |
| Simulation | Number of Orders in the Backlog | Two heaps as a simple order book |
| Implementation speed | Two mediums back to back in 35 minutes | Clean, first-time-correct code |
HRT's test is coding-first. Prep guides stress probability and mental math for algo developer and quant-leaning roles, but we found no consistent report of a separate timed arithmetic test in HRT's online assessment like those some market makers use. Expect that material in later interviews instead.
Scoring, pass bar and proctoring
- No published pass mark. HRT does not publish a cutoff, and a perfect score is not documented as required. At a firm this selective, assume partial credit on the hard problems matters.
- Speed and exactness. Candidates describe solutions that must be both correct on hidden tests and fast enough not to time out.
- Proctoring. If your invitation is a proctored CodeSignal test, expect an identity check, webcam and screen sharing; HackerRank tests can track tab switches and copy-paste. HRT's postings say it checks the authenticity of responses. Our HackerRank assessment guide covers that platform's monitoring.
- Language. Use the fastest language the test allows. If it is C++ only, practice C++ for the whole plan.
A 14-day preparation plan
Two weeks assumes you already solve LeetCode mediums reliably. If you do not, spend an extra week on mediums before starting.
- Days 1–3: Graph algorithms. Dijkstra, 0-1 BFS, union-find and topological sort, three problems a day.
- Days 4–6: Dynamic programming. Interval, two-sequence and bitmask DP; write the state and recurrence before any code.
- Days 7–8: Strings and number theory. Sliding windows, sieves, fast exponentiation and modular inverses.
- Day 9: Simulation and heaps. Order-book style problems and event processing.
- Days 10–11: Speed sets. Four problems in 70 minutes, escalating from easy to hard, twice a day.
- Day 12: Full simulation. An unseen four-problem set under the reported time, no notes, your own tests.
- Day 13: Review. Redo every timeout and wrong answer, and note the bug patterns.
- Day 14: Light work and logistics. One easy warm-up, an equipment check and a careful read of the invitation's rules.
What happens after the assessment
HRT's own description lists about two phone interviews, then a full day of interviews. Candidates describe live coding in a shared editor, systems questions on memory, concurrency and Linux, and for algo roles probability and statistics. Our Hudson River Trading coding interview questions cover those rounds, and the quant interview prep guide covers the probability side. To compare first screens across firms, see the Citadel online assessment and the Two Sigma online assessment.
Where CoPilot Interview fits
CoPilot Interview is an AI interview assistant, and for the HRT online assessment its only honest role is practice beforehand. The test measures your own skills under the rules in your invitation, so no assistant, AI tool or second person belongs in it unless the invitation explicitly allows one. Before the test, the browser demo has 145 interview questions across 10 tracks, scores typed answers out of 10 and needs no signup. Once you pass, the rounds become spoken, and the AI mock interview reads questions aloud and scores each spoken answer out of 10, which is the way to rehearse explaining your approach out loud. The free plan gives 15 minutes a day of live help in sessions of up to 15 minutes, with AI mock interviews included. Standard costs $24.99 a month and Pro $39.99 a month if you want more practice time. For every live round later in the process, follow your employer's rules. HRT's postings say AI tools are prohibited in its interviews and assessments unless you are told otherwise.
Disclosure: CoPilot Interview is our product, so weigh what we say about it accordingly.
Frequently asked questions
What platform does HRT use for its online assessment?
Recent software engineering intern candidates most often report CodeSignal, and others report HackerRank. HRT's 2021 blog post named HackerRank or Codility. Your invitation names the platform.
How many questions are on the HRT OA?
Recent software engineering reports most often describe four questions in about 70 minutes, and others describe three or four problems in 70 to 90 minutes. Algo developer candidates report a longer or harder set.
Do I need a perfect score to pass the HRT OA?
HRT publishes no cutoff, and a perfect score is not documented as required. Aim to fully solve the earlier problems and earn partial credit on the hardest.
Is there a math test in the HRT assessment?
Not consistently. HRT's online assessment is coding-first; probability and statistics are reported mainly in later interviews for algo developer and quant-leaning roles.
Can I use AI tools or the internet during the HRT assessment?
Not AI. HRT job postings say AI tools are strictly prohibited during interviews and assessments unless you are instructed otherwise. A 2021 HRT post allowed books and the internet as a language reference, but follow your current invitation's rules, not that post.
Practice this interview before the real one
The AI mock interview builds questions from your resume and the job description, reads them aloud and scores each spoken answer out of 10. Included on the free plan.
Start a mock interview