XTX Markets Interview Guide 2026: ML-Driven HFT, FX Dominance, London-Headquartered Quant Powerhouse

XTX Markets Interview Guide 2026: ML-Driven HFT, FX Dominance, London-Headquartered Quant Powerhouse

XTX Markets is one of the most distinctive HFT firms globally — founded in 2015 by Alex Gerko (former Deutsche Bank statistical arbitrage trader), the firm rapidly rose to dominate FX market making and has expanded into equities, fixed income, and commodities. Unlike most US-headquartered HFT firms, XTX is London-based with a distinctive ML-first approach (the firm reportedly runs the largest ML training infrastructure of any non-tech-giant in finance). XTX has been one of the biggest hedge fund / HFT philanthropic donors globally, with Alex Gerko publicly committing to donate the majority of his wealth. The hiring process is rigorous and reflects the firm’s research-driven culture. This guide covers what XTX does, the engineering tracks, the interview process, and what makes XTX hiring distinctive in 2026.

What XTX Does

XTX operates as a quantitative trading firm:

• FX market making: the firm’s flagship business. XTX is among the top three FX market makers globally by volume, alongside Citadel Securities and Jump Trading. Substantial market share in EBS and Refinitiv liquidity pools.
• Equities market making: US equities, European equities. Growing post-2020.
• Fixed income: rates, credit market making at growing scale.
• Commodities: energy and metals trading.
• Crypto: XTX has reportedly entered crypto market making in some capacities; less transparent than the FX / equities business.
• Research and ML infrastructure: XTX runs the largest publicly-known ML training infrastructure in finance (reported to have 25,000+ NVIDIA H100/H200 GPUs as of 2024). The firm’s research uses ML models for signal generation across asset classes.

Distinctive features:

• ML-first approach: XTX’s competitive edge is reportedly ML-driven signal extraction and execution. Substantial GPU infrastructure investment differentiates XTX from peer HFT firms that use simpler statistical models.
• London headquarters: distinct from US-headquartered HFT peers. UK-based engineering culture, regulated by FCA. Singapore and Mumbai offices for research and engineering.
• Privately held: Alex Gerko owns a substantial stake; never raised external capital. Public profitability disclosures from UK reporting requirements show XTX as one of the most profitable financial firms globally.
• Philanthropic commitment: Alex Gerko has signed The Giving Pledge and is one of the largest finance-industry philanthropists globally. The firm’s culture reflects this somewhat — less aggressive than typical HFT firms in commercial messaging.
• Research-paper publication: XTX has published academic papers (especially in ML and signal processing) and supports academic conferences. Less common among HFT firms historically.

Roles XTX Hires For

Quantitative researcher

Develops trading strategies, ML signal models, execution algorithms. Deep ML / statistical learning expertise. PhD-friendly but exceptional MS candidates considered. Substantial recent hiring growth as XTX expands into new asset classes.

Software engineer (low-latency / trading systems)

Builds the core trading infrastructure — order management, execution, risk systems, latency-critical code paths. C++ heavy; some Rust adoption. Performance engineering deep.

Software engineer (ML infrastructure)

Builds and maintains XTX’s substantial ML training infrastructure — GPU clusters, distributed training systems, data pipelines for ML feature generation. Distinctive at HFT firms; XTX’s ML infra team is reportedly larger than most peer HFT firms’ entire engineering departments.

Data engineer

Pipeline engineering for the substantial market data and alternative data XTX consumes. Tick data, order book data, alternative data ingestion at scale.

Network / systems engineer

Latency-critical network engineering — co-location strategies, hardware optimization, fiber routing decisions. Specialized work.

Research engineer (bridge between research and production)

Productionizes research models — taking strategies from research environments to production trading systems. Hybrid of ML engineering and trading systems.

Risk / portfolio engineer

Real-time risk monitoring, portfolio analytics, stress testing across asset classes. Substantial work given XTX’s multi-asset trading footprint.

XTX Interview Process

Round 1: Recruiter screen

30 minutes. Background, motivation, role fit. Recruiters often probe research / ML engagement specifically.

Round 2: Online assessment / take-home

For research roles: a take-home statistical / ML problem with a 24–48 hour window. Typically open-ended (build a model that predicts X) with the firm evaluating both technical quality and reasoning. For engineering roles: HackerRank-style timed coding test.

Round 3: Technical phone / video screen

60–90 minutes. For research: discuss the take-home, probability questions, ML theory. For engineering: coding plus systems / latency questions.

Round 4: On-site / virtual on-site

4–6 rounds, each 60–90 minutes:

• Coding (1–2 rounds) — algorithms with practical engineering flavor; C++ depth tested for systems roles
• Research / ML deep dive (1 round for research roles) — discuss your work, defend methodology, engage with critique
• Probability / statistics (1 round) — fundamental probability, statistical inference, ML theory
• System design (1 round for engineering roles) — distributed systems, low-latency architectures
• Behavioral / cultural fit (1 round) — collaboration, intellectual humility, firm-fit

Round 5: Decision

Calibration meeting; offer typically within 1–3 weeks. Compensation negotiation expected.

What XTX Tests For

ML / statistical depth

For research roles, XTX expects deep ML expertise — beyond surface-level model usage. Understanding of statistical foundations, regularization, optimization, distributional shifts, model evaluation methodology. Engineers from data science backgrounds need to demonstrate methodological rigor.

Quantitative reasoning under pressure

Probability questions and statistical reasoning matter. Strong candidates can discuss conditional probability, hypothesis testing, regression diagnostics, time series stationarity at depth.

C++ depth (for engineering roles)

Most XTX systems are C++ at depth. Memory management, template metaprogramming, performance optimization, lock-free programming all matter. Engineers from Java / Python backgrounds need substantial C++ ramp.

Research engineering bridge skills

For research engineer roles, ability to bridge between ML research and production trading systems. Understanding of how models behave in production, latency constraints, monitoring requirements.

Cultural fit — intellectual but humble

XTX’s culture rewards intellectual depth without arrogance. Candidates who oversell their work or refuse to engage with critique underperform. The “can you say I don’t know” test matters.

Compensation

Among the highest in HFT for senior roles; competitive but generally lower for junior roles than top US-based prop firms:

• New-grad QR / SWE: £100k–£200k total comp first year (London market)
• Mid-level (4–7 years): £200k–£500k
• Senior (8+ years): £500k–£2M+
• Senior researcher / Principal: £1.5M–£10M+ for top performers

Compensation is base + bonus, with bonus heavily performance-driven. UK tax structure differs from US substantially; UK income tax tops out around 47% above £150k threshold. Senior staff often relocate to lower-tax jurisdictions when possible.

For US-based hires (smaller team), compensation is broadly comparable to top US HFT firms ($350k–$3M+ range across career stages).

Working at XTX

Tech stack and engineering quality

C++ heavy for trading systems; Python for ML and research; some Rust adoption for newer projects; substantial CUDA / ML framework work for the ML infrastructure team. Engineering quality is regarded as high; the ML infrastructure investment is genuinely distinctive in finance.

Pace and intensity

Moderate-to-intense. Less frenetic than US pod-shop hedge funds; more focused than typical UK financial services. London office work culture is generally more sustainable than NYC peers (statutory holiday entitlement, parental leave, etc.).

Office and remote

HQ in London (Mayfair / St James’s area). Major offices in Singapore (Asia trading and research), Mumbai (engineering and research), New York (smaller US office). Hybrid model post-COVID; substantial in-office expectation given XTX’s collaborative research culture.

Career trajectory

Standard quant-firm leveling. Long tenures common — XTX has retained senior researchers and engineers from its early days. Promotion is rigorous; the firm has a reputation for high bars and selective progression.

XTX vs Alternatives

XTX vs Citadel Securities: Both top market makers. Citadel Securities is broader (US equities + options + FX + futures + Treasuries); XTX is FX-dominant with growing equities. Citadel Securities is US-headquartered with NYC-style culture; XTX is London-headquartered with European style. Compensation comparable at senior levels.

XTX vs Jump Trading: Both ML-aggressive HFT firms. Jump is Chicago-based; XTX is London-based. Jump’s crypto exposure (via the Jump Crypto subsidiary) is substantial; XTX’s crypto exposure is more limited. Engineering culture similar — both research-driven and ML-investment-heavy.

XTX vs Jane Street: Both top quant firms. Jane Street is US-headquartered, OCaml-flavored, options-and-equities focused; XTX is London-based, ML/CUDA-flavored, FX-and-multi-asset focused. Different engineering cultures; both are top destinations for quant talent.

XTX vs Optiver: Both have substantial European footprint (Optiver is Amsterdam-headquartered, XTX is London). Optiver is options-market-making focused; XTX is FX-and-multi-asset. Different specialty depths.

Things That Surprise Candidates

• The ML infrastructure investment is more substantial than candidates expect; the GPU footprint is in the same league as some AI labs.
• The research culture is closer to academic AI labs than typical HFT firms — published papers, conference engagement, exploratory research time.
• The London-NYC tax differential is real; engineers calibrating offers should compute net comp carefully.
• The privately-held / no-investor-pressure structure means longer-horizon thinking than at most peers.
• The philanthropic commitment is genuine — Alex Gerko’s public stance influences firm culture, though most engineers aren’t directly involved in firm philanthropy.

Frequently Asked Questions

See also: Jane Street Interview Guide • Breaking Into Quant Finance • C++ for Quant Interviews