GSA Capital hires Quant Researcher and Quant Developer. Candidates report 4 to 5 interview stages depending on the track (5 distinct rounds across all tracks). Below: what each round tests and how to prepare, the path by track, the timeline, pay and a 4-week prep plan, drawn from candidate reports and GSA Capital's own recruiting pages.
GSA Capital interview: quick answer
- StagesBy track: Quant Researcher: 5 · Quant Developer: 4 (5 distinct rounds across all tracks)
- Online assessmentCorrelation One · 25 questions / 60 min · MCQ + numeric entry
- TimelineGSA's published process has no fixed length, and the firm says it is not the shortest.
- Final roundFinal: Conversational · with senior researchers
- Pay (candidate-reported)Quant Researcher, New grad / PhD: $250k – $450k total comp
- RolesQuant Researcher, Quant Developer
Preparing for GSA Capital? Work the rounds in order in the free GSA Capital interview funnel, or start with the free online-assessment sample (no signup).
The GSA Capital interview process at a glance
| # | Round | Format and length | Who takes it | Prep |
|---|---|---|---|---|
| 1 | Online Assessment Correlation One | Correlation One · 25 questions / 60 min · MCQ + numeric entry | Quant Researcher | OA guidePractice |
| 2 | Take-Home Data Project | Take-home · own environment · graded on OOS MSE | All tracks | GuidePractice |
| 3 | Technical Regression & Statistics | 45–60 min · with a researcher | All tracks | GuidePractice |
| 4 | Portfolio Construction & Machine Learning | 45–60 min · with a researcher | All tracks | GuidePractice |
| 5 | Final Research Fit | Conversational · with senior researchers | All tracks | GuidePractice |
Stages are listed in the order candidates report them. Formats come from candidate reports and change between cycles, offices and teams; the "Who takes it" column shows which track each round belongs to.
How the GSA Capital process differs by track
GSA Capital runs separate loops for different roles. Here is the path each track reports, in order:
- Quant Researcher (5 stages): Online Assessment → Take-Home Data Project → Technical → Portfolio Construction & Machine Learning → Final
- Quant Developer (4 stages): Take-Home Data Project → Technical → Portfolio Construction & Machine Learning → Final
Not sure which track fits you? See quant trader vs quant researcher.
Every GSA Capital round: what it tests and how to prepare
1Online Assessment: Correlation One
Quant Researcher first screen, Correlation One: 25 questions in 60 minutes, single-option MCQ or numeric entry, PREVIOUS / NEXT navigation, 25-box progress bar. OA page: timed replica, verified answers.
Aug 2026 sitting, four families. ML/statistics: k-means local optima, overfitting, residual-vs-one-variable plots, R-squared vs t-statistics, regression on 110 observations, 100 variables.
Brainteasers: nine dots in four segments, molecule-collision invariant, knights and knaves.
How to prepare
- Learn the numeric-entry conventions. Probability answers are typed as the integer closest to 20p and fractions as M + N for M/N in lowest terms; do the arithmetic before you move on because numeric items get no partial credit.
- Know the ML concept traps. Restarts (not random-data-point initialization) fix k-means local optima; more features worsens overfitting; one-variable residual plots miss interactions; high R-squared and significant t-statistics do not imply each other; with n about p reach for lasso/ridge, not OLS or best subsets.
- Use the navigation. You can go back, so bank the quick concept and judgment items first and return to the numeric ones (eigenvalue sums, nullity, integrals) with the time you saved; 60 minutes for 25 questions is about 2.4 minutes each.
2Take-Home Data Project
GSA's signature stage. From real 10-second-bar prices for instruments X and Y, predict Y's 10-minute-ahead return with data up to t only, in two template functions.
- GSA's advice: keep it simple, and pay attention to design, tests, documentation and maintainability.
- A recruitment alignment call happens while you work on it. Use it to clear up anything ambiguous in the brief.
How to prepare
- Respect causality. Only use information available at time t; any feature that peeks at the future inflates in-sample fit and tanks OOS MSE.
- Drop degenerate labels. The last ~10 minutes of each trading day have truncated forward returns — exclude them from training.
3Technical: Regression & Statistics
The core technical round: regression theory and statistics. Expect ridge closed form and why X'X + λI is invertible, variance of a sum of equicorrelated variables (and valid-ρ bounds), the bus-stop inspection paradox, years-until-positive-return via the CLT, Sharpe-ratio backtest length, and estimator reasoning. They want a clean argument, not just a number.
- GSA's process opens with two remote technical interviews of up to an hour on foundations in your field, and later has a second round of up to about four hours of team-specific interviews.
Question types: mostly probability, statistics and expectation, especially expected value, normal distribution, conditional probability and symmetry. Latest candidate report for this round: Jul 2026.
How to prepare
- Re-derive ridge and OLS from scratch. Know β̂_ridge = (X'X + λI)⁻¹X'Y and why adding λI > 0 forces positive-definiteness (hence invertibility).
- Know length-biasing. The inspection paradox and equicorrelation bound ρ ≥ −1/(n−1) are recurring.
4Portfolio Construction & Machine Learning
A round on portfolio optimization and, increasingly, machine learning. Expect Markowitz optimization given an alpha signal and covariance matrix (objective, constraints, unconstrained closed form), covariance-estimation methods (sample vs EWMA vs PCA factor models vs graphical Lasso), and recent ML items: PCA components with near-zero R², tree feature importance / SHAP values, and neural-net initialization / batch-norm-and-dropout at test time.
- This content belongs to the second round, up to about four hours of deeper, team-specific interviews.
Question types: mostly machine learning, statistics and optimization, especially regression, feature engineering, model selection and regularization. Latest candidate report for this round: Jul 2026.
How to prepare
- Set up mean-variance cleanly. Maximize w'μ − (γ/2)w'Σw; unconstrained gives w* ∝ (1/γ)Σ⁻¹μ; know the sum-to-one Lagrange version.
- Have a covariance-estimation menu. Sample covariance breaks when N ≫ T; be ready to compare EWMA, PCA factor models, and graphical Lasso (precision-matrix MLE).
- Connect ML to linearity. Low R² on top PCA components often just means the signal is nonlinear; know that a univariate SHAP value reduces to the regression beta.
Full Portfolio Construction & Machine Learning guide →Practice this round (8 questions) →
5Final: Research Fit
A closing research-fit stage: a deep-dive on your take-home methodology and past projects, plus culture signal. They probe why you made each modeling choice, how you validated it, and how you'd extend it.
- Final conversations with senior or technical leaders focus on long-term potential and fit across teams, and the route can shift to a different team.
How to prepare
- Defend every take-home choice. Feature selection, horizon, validation scheme — treat the deep-dive like a research review.
- Be candid about limitations. Naming what you'd do with more time reads as maturity, not weakness.
Practice the GSA Capital rounds in order
Each round in the GSA Capital funnel has its own practice set, with worked solutions and progress tracking. A free account saves your progress across rounds.
Timeline
GSA's published process has no fixed length, and the firm says it is not the shortest. Formats can vary by team.
- Two remote first-round technical interviews of up to an hour each, on HackerRank or Zoom.
- A take-home exercise in your own time, with a recruitment alignment call while you work on it.
- A second round of up to about four hours of deeper, team-specific interviews.
- Final conversations with senior or technical leaders.
A 4-week GSA Capital prep plan
Built from the order of GSA Capital's rounds on the Quant Researcher path and the topics in each round's practice set. Compress it if your first interview is sooner. Other tracks can swap in their own rounds from the table above.
| When | Focus | What to do |
|---|---|---|
| Week 1 | Foundations | Review statistics, machine learning and probability, the topics that dominate GSA Capital's practice sets. Start with Statistics & regression course and Machine learning course. Work the easy and medium questions in the GSA Capital question set. |
| Week 2 | Early rounds: Online Assessment | Practice the screening topics under time pressure, and take the free OA sample (format notes in the OA guide). Add a daily mental-math drill. Talk every answer through out loud; interviewers score the reasoning. |
| Week 3 | Later rounds: Take-Home Data Project, Technical, Portfolio Construction & Machine Learning and Final | Move to statistics, machine learning and probability. Read how quant superdays work. |
| Week 4 | Rehearse the full loop | Run each practice round in order, redo the questions you missed, and prepare two project stories plus a specific answer to "why GSA Capital?" (see behavioral questions at trading firms). |
GSA Capital pay
| Role | Level | Total comp | Notes |
|---|---|---|---|
| Quant Researcher | New grad / PhD | $250k – $450k | Base + performance bonus; London package. Candidate-reported. |
| Quant Developer | New grad | $200k – $350k | Engineering ladder alongside the research bar. Candidate-reported. |
Total compensation means base salary plus bonus and any sign-on, as candidates self-report it. Job postings usually show base salary only, so posted figures and outside salary pages often look lower. Rough, candidate-reported ranges — NOT offers. GSA is a London systematic fund; comp is heavily bonus-driven and varies with performance and role. Sanity-check only. Source: Candidate self-reports (levels.fyi / Glassdoor / community forums) — not employer-verified. Last updated July 17, 2026.
GSA Capital teams and what their interviews test
GSA hires quant researchers and developers into a systematic statistical-arbitrage platform. Researchers own signal and portfolio-construction work; the interview bar opens with a Correlation One online assessment and centers on regression theory, statistics, and a real modeling take-home.
| Tracks | What it does | What its interviews test |
|---|---|---|
| Quant Researcher Quant Researcher | Signal research, estimation, and portfolio construction on systematic strategies. | Correlation One OA (25 q / 60 min: ML-stats concepts, numeric-entry probability, brainteasers, judgment items) → take-home data project → regression / statistics / Sharpe-and-backtest reasoning → portfolio optimization + ML → research fit. Rigor about estimation and validation is the differentiator. |
| Quant Developer Quant Developer | Research and trading infrastructure supporting the systematic platform. | Coding + systems alongside the shared statistics bar; clean, correct implementation under the same research-first culture. |
About GSA Capital
| Firm type | Systematic single-manager |
|---|---|
| Trading style | Systematic / statistical arbitrage · Mid- to high-frequency |
| Founded | 2005 |
| Headquarters | London, UK |
| Other offices | New York |
| Assets under management | ~$7–8B (reported) |
| Employees | ~150 (candidate-reported) |
| Published process | Two remote technical interviews, a take-home with an alignment call, a second round of up to about 4 hours, then final conversations with senior leaders |
| Size | About 150 people, per GSA |
Roots in Deutsche Bank statistical arbitrage
GSA Capital is a London-based systematic hedge fund that traces its origins to Deutsche Bank's Global Statistical Arbitrage group before becoming an independent manager. Its DNA is quantitative statistical-arbitrage research.
Culture
Research-heavy systematic shop with a strong statistics / regression bar — GSA grew out of a bank statistical-arbitrage desk and hires researchers who can reason rigorously about signals, estimation, and portfolio construction.
What the interviews feel like
A 25-question, 60-minute Correlation One online assessment (ML/stats concepts, numeric-entry probability, brainteasers, cognitive-bias judgment items) opens the researcher funnel; a real take-home data challenge (10-minute-ahead return prediction, graded on out-of-sample MSE) anchors the loop, surrounded by technical rounds on regression theory, statistics, Sharpe / backtest reasoning, portfolio optimization, and increasingly machine learning.
Practice for GSA Capital on QuantVault
Practice
Courses
Games and drills
Further reading
- Behavioral questions at trading firms
- Mental math for trading interviews
- Fermi estimation questions
- How quant superdays work
- Market-making game interview guide
- The quant interview process, explained
Similar firms to compare
Firms with a similar business and interview style. Many candidates interview at several of them in the same cycle.
Sources and how this guide is built
- Updated October 20, 2026; firm data last reviewed September 6, 2026; pay last updated July 17, 2026.
- Round detail draws on 12 questions candidates report from GSA Capital interviews (the most recent report across all rounds is from Jul 2026). A separate curated GSA Capital question set of 26 problems collects the most useful practice for this firm.
- GSA Capital's own pages: Our interview process · GSA Capital.
- Public candidate reports and outside interview guides are cross-checked against each other. Where they disagree the page says "reports differ"; details from a single source are marked as candidate reports.
- No actual interview questions are reproduced on this page. The questions, with worked solutions, are in the practice rounds.
GSA Capital interview FAQ
How many rounds are in the GSA Capital interview?
It depends on the track: Quant Researcher: 5; Quant Developer: 4 stages. Across all tracks candidates report 5 distinct rounds: Online Assessment, Take-Home Data Project, Technical, Portfolio Construction & Machine Learning, Final.
How long does the GSA Capital interview process take?
GSA's published process has no fixed length, and the firm says it is not the shortest.
What happens in the GSA Capital final round?
Final (Conversational · with senior researchers). A closing research-fit stage: a deep-dive on your take-home methodology and past projects, plus culture signal. They probe why you made each modeling choice, how you validated it, and how you'd extend it.
Does the GSA Capital process differ by role?
Yes. Quant Researcher: Online Assessment, Take-Home Data Project, Technical, Portfolio Construction & Machine Learning, Final. Quant Developer: Take-Home Data Project, Technical, Portfolio Construction & Machine Learning, Final.
How much does GSA Capital pay?
Candidate-reported total compensation: Quant Researcher (New grad / PhD) $250k – $450k; Quant Developer (New grad) $200k – $350k. These are self-reported totals including bonus, not offers; job postings usually list base salary only.
How should I prepare for the GSA Capital interview?
Prepare round by round over about four weeks: foundations first, then the early rounds (Online Assessment, Take-Home Data Project, Technical), then the later rounds, then a full rehearsal. Each round has its own practice set on QuantVault, and the online assessment has a dedicated guide.
What does GSA Capital do?
GSA Capital is a systematic single-manager headquartered in London, UK. GSA Capital is a London-based systematic hedge fund that traces its origins to Deutsche Bank's Global Statistical Arbitrage group before becoming an independent manager. Its DNA is quantitative statistical-arbitrage research.
Is there an online assessment at GSA?
GSA's published process starts with remote technical interviews and lists no separate test. Quant research candidates in August 2026 reported a Correlation One assessment, which may be specific to that track.
Do I need finance knowledge for a software role at GSA?
No. GSA says software engineers do not need a finance background.
Start preparing for GSA Capital
Open the interactive GSA Capital interview funnel: every stage with its practice set, from the first screen to the final round. Create a free account to save your progress.