Hudson River Trading (HRT) Interview Guide

How the HRT interview process works

HRT publishes a four-step student recruiting process. For engineering candidates, the phone stage typically includes a technical discussion and a programming interview, followed by a full day of final-round sessions. Here's the usual progression:

1. Application review. Your background and the role you applied for determine whether HRT invites you to the next stage.
2. Online coding assessment. Solve timed implementation and algorithm problems, then check hidden-test edge cases.
3. Technical phone screens. Expect live programming, plus a roughly 45-minute engineering discussion or probability-and-coding work for Algorithm Developers.
4. Final-round interviews. Three to five sessions test deeper technical work, design and collaboration.

The roles make the biggest difference. HRT calls much of its mathematical research and trading work Algorithm Development. That is a different track from Software Engineering, even though both require strong programming skills. We'll cover the shared stages first and then the preparation needed for each.

Application and initial screening

HRT asks students to apply to one job posting, with other office or role interests noted in the application. Allow up to two weeks for application review. The recruiting team then emails selected applicants with assessment or interview instructions.

Match your application to the actual job. An Algorithm Developer application should make your statistics, research, programming and analytical projects easy to find. For Software Engineering, put substantial engineering work near the top: what you built, the languages you used, and the technical decisions you made. HRT also advises experienced applicants to start with the role that best matches their strengths.

You don't need previous financial-market experience to apply. HRT explicitly welcomes candidates without it. A strong technical foundation is more relevant at this stage than memorising trading terminology.

The HRT online assessment and coding test

The HRT online assessment (OA) is a timed programming test. You write working solutions, run sample cases and submit code against hidden tests. The engineering guidance puts particular emphasis on correctness and edge cases, even when the visible examples all pass.

Prepare for arrays, strings, hash maps, grids, interval problems and multi-step simulations. In a July 2026 new graduate report, two OA tasks involved finding the largest diamond-shaped sum in a grid and the position covered by the most street lamps. Both reward careful implementation, efficient counting and boundary tests.

HackerRank, Codility or CodeSignal?

CodeSignal appears in recent Software Engineering applicant reports, including a 2025 new graduate assessment taken in C++. HRT's engineering recruiting guide also names HackerRank and Codility for take-home coding tests. All three reward executable code that handles inputs beyond the samples.

How many questions, and what score do you need?

For a recent Software Engineering CodeSignal OA, four questions in 70 minutes is the clearest reported format. The early problems can be short, while later ones demand longer implementations. A September 2026 Algorithm Developer applicant also reported a 70-minute coding OA. There is no verified numerical HRT passing cutoff, so practise finishing working solutions and passing every test you can.

How to practise for the OA

Do a four-problem, 70-minute coding mock in your chosen language. Include a quick array task, a hash-map or interval problem, a grid problem and a longer stateful simulation. Test empty inputs, repeated values, off-by-one boundaries and large cases. For event streams, verify the state after each update, not just the final answer.

Technical phone and video interviews

HRT's published Software Engineering process has two types of calls: an approximately 45-minute discussion on systems, data structures or problem-solving, and a live programming interview. Recent candidates also describe shorter 30-minute coding screens. Expect to work in a shared editor, explain the algorithm, and improve it when the interviewer adds requirements.

For Algorithm Developers, 2026 reports describe fast-paced probability and expected-value questions with numerical answers, followed by coding or a Python simulation. One candidate reported two maths questions and a two-part coding task. Software Engineers are more likely to explain memory, data structures, program behaviour and debugging, including what their chosen language does under the hood.

Use your explanations to make the work understandable. State your assumptions, outline an approach, and say what you've tested. If you spot a mistake, correct it and continue. HRT's own interviewers say they value candidates who can work through guidance and improve a solution.

What happens in the final-round interview?

The HRT onsite or final round typically consists of three to five interviews. Its engineering guidance describes a full day of back-to-back coding and debugging, technical design, and team conversations. These sessions test whether you can build and reason through working solutions, not just finish short screening questions.

For Software Engineering, prepare to discuss memory and I/O, diagnose broken code and defend performance trade-offs. A reported Algorithm Developer Superday included dataset analysis, a codebase exercise, system design and a behavioural interview. That is a useful model for practising both open-ended research and hands-on coding.

Be prepared to discuss something you've built. You should be able to explain why you chose a particular approach, what went wrong, how you tested it and what you would change if you started again. Real project details give interviewers a better basis for follow-up questions than a rehearsed summary does.

Interview questions by HRT role

Choose the track closest to your application. HRT's job titles matter here: its Algorithm Developers research and improve trading models, while Software Engineers and Algorithm Engineers build the infrastructure those models run on. A specialised Quantitative Researcher or Data Science role may have a different technical emphasis.

HRT-style technical questions, with solutions

These are original practice examples chosen to exercise skills relevant to HRT interviews. Try solving each one before reading the explanation.

def net_positions(events):
    active = {}
    positions = {}

    for event in events:
        if event[0] == "fill":
            _, fill_id, symbol, qty = event
            active[fill_id] = (symbol, qty)
            positions[symbol] = positions.get(symbol, 0) + qty
        else:
            _, fill_id = event
            original = active.pop(fill_id, None)
            if original is not None:
                symbol, qty = original
                positions[symbol] -= qty

    return {s: q for s, q in positions.items() if q != 0}

The first question tests whether you can turn a repeated random experiment into a simple equation. The second tests careful state management and edge cases. Those are useful habits to develop for both a coding screen and a more open-ended technical interview.

How difficult is the HRT interview?

HRT is a high-difficulty technical interview. On the reported 70-minute, four-question Software Engineering OA, speed matters as much as recognising an algorithm: later problems can require grid calculations, interval optimisation or detailed simulation. Algorithm Developer screens add the pressure of solving probability questions and then implementing checks under live interview conditions.

HRT's engineering team says it prefers practical coding over trick-heavy LeetCode puzzles. The harder part is often explaining fundamentals precisely: how Python lists grow, how C++ virtual calls work, or why a solution behaves differently on large inputs. Write correct code, test corner cases and be prepared to explain its resource costs.

For Algorithm Developers, treat statistics and programming as equally serious parts of the process. For Software Engineers, prioritise code quality, debugging and the systems relevant to your role.

How to prepare for the HRT interview

Prepare for the next round you actually face. For a Software Engineering OA, rehearse four tasks in 70 minutes. For an Algorithm Developer phone screen, pair expected-value problems with short simulations. For a C++ infrastructure final, spend more time on memory, debugging and system behaviour.

1. Rehearse the coding test. Solve four problems in 70 minutes: arrays or strings, hash maps or intervals, grids, and stateful simulation. Test hidden-case boundaries yourself.
2. Study your role's deeper topics. For Algorithm Development, practise conditional probability, expected value, estimation and Python Monte Carlo. For Software Engineering, explain data structures, virtual memory, C++ dispatch or Python runtime behaviour.
3. Practise a live screen. Spend 45 minutes explaining a coding or maths solution aloud while someone asks follow-ups. Prepare a clear account of one difficult project.
4. Revisit your mistakes. Keep a short record of errors in reasoning, edge cases and code. Re-solve those questions until the fixes feel routine.

For focused practice, the EverythingQuant HRT question bank targets probability and stochastic reasoning. If expected value or conditional probability needs work, review the Probability for Quants course. Use the coding questions for implementation and the DSA course to strengthen fundamentals. If you're preparing across several firms and assessments, you can also build a personalised preparation roadmap.

HRT interview FAQs