How to Get a Job at XTX Markets

XTX Markets is a London-based algorithmic trading firm founded in 2015 by Alex Sherbrooke as a spinoff from GSA Capital. Despite its relative youth, XTX has rapidly become one of the largest market makers in the world, consistently ranking among the top 5 in global FX and equities. The firm trades across equities, fixed income, FX, and commodities on over 100 exchanges, providing liquidity through fully automated, algorithm-driven systems.

What makes XTX unique in the trading industry is its purely systematic approach. Unlike firms that employ human traders to make discretionary decisions, XTX relies entirely on machine learning models and algorithms to price, execute, and manage risk. There are no human traders on the desk — every trading decision is made by the firm's models. This makes XTX more of a technology and research company that happens to trade, rather than a traditional trading firm.

XTX manages this through a combination of cutting-edge machine learning research and massive engineering infrastructure. The firm processes enormous volumes of market data, applies sophisticated statistical models to identify pricing relationships, and executes trades with minimal latency. With approximately 200 employees — most of them researchers and engineers — XTX operates with an exceptionally high revenue-per-employee ratio. The firm's London headquarters is supplemented by offices in Singapore and Mumbai, and it has been profitable every year since inception.

XTX's culture reflects its identity as a research-driven technology company. The firm attracts people who are passionate about machine learning, statistics, and building systems at scale. The atmosphere is more academic than most trading firms — employees spend their time reading papers, implementing algorithms, running experiments, and debating research approaches rather than watching markets tick by tick.

The firm operates with a relatively flat structure and small team size. With around 200 people generating billions in revenue, every employee's contribution is visible and impactful. Teams are organized around research areas (signals, execution, infrastructure) and collaboration is essential — researchers need engineers to productionize their models, and engineers need researchers to guide system design. This interdependence creates a tight-knit environment where mutual respect across disciplines is the norm.

XTX is also known for its commitment to social impact. The firm and its founder have donated hundreds of millions of pounds to charitable causes, particularly in AI safety, mathematics education, and scientific research. This philanthropic orientation attracts people who want their work to have broader positive impact. Work-life balance at XTX is generally better than at many competitors — the absence of human traders means there is less pressure tied to market hours, and the firm values sustainable long-term productivity over intense short-term sprints.

XTX hires primarily for two types of roles: quantitative researchers (who build the models that drive trading) and engineers (who build the systems that execute those models at scale). For both, the bar is exceptionally high — XTX is selective about hiring people who are genuine experts in their domains.

For research roles, XTX looks for candidates with deep machine learning expertise — not just familiarity with scikit-learn, but genuine understanding of modern ML methods including deep learning, reinforcement learning, Bayesian methods, and causal inference. Many researchers have PhDs in machine learning, statistics, or physics, though exceptional candidates with master's degrees or strong industry experience are also considered. Publications in top ML venues (NeurIPS, ICML, JMLR) are a strong positive signal.

For engineering roles, XTX seeks candidates with strong systems programming skills — typically C++ and Python, with deep understanding of performance optimization, distributed systems, and data infrastructure. The firm's trading systems process enormous data volumes with stringent latency requirements, so engineers need to be comfortable operating at the intersection of software engineering and systems design. Experience with large-scale ML infrastructure, real-time data processing, or high-performance computing is particularly valued. Across all roles, XTX wants people who are intellectually rigorous, genuinely curious about how markets work, and excited to solve hard problems at the intersection of ML and finance.

XTX's interview process typically consists of 4 to 5 rounds over 4 to 8 weeks. The process is heavily focused on technical depth — XTX wants to verify that candidates have genuine expertise, not just surface-level familiarity with the relevant topics. The firm is known for asking challenging, open-ended research questions that test depth of understanding.

The general structure is:

• Initial screen (1 round): A phone or video call with a team member, typically lasting 45-60 minutes. This covers your background, research experience, and includes some technical questions tailored to your area (ML for researchers, systems design for engineers). The goal is to assess whether you have the baseline expertise for a deeper evaluation.
• Technical assessment (1 round): For researchers, this may be a take-home problem involving data analysis, model building, or a theoretical ML question. For engineers, it may be a coding challenge focused on systems design or performance optimization. The assessment tests your ability to produce high-quality work independently.
• On-site interviews (2-3 rounds): Deep technical discussions with multiple team members. Researchers face questions on ML theory, experimental design, and applied statistics. Engineers face systems design and coding challenges. Both may discuss how they would approach real trading problems.
• Final round: A conversation with senior leadership or the team lead, focusing on research direction, long-term goals, and cultural fit.

XTX's interviews are distinctive in their depth over breadth — rather than covering many topics superficially, interviewers will pick one or two areas and probe deeply to understand the limits of your knowledge. Being honest about what you don't know is valued over bluffing.

Each round of the XTX interview tests specific aspects of your expertise. Here's what to expect:

Machine Learning Theory and Practice (Researchers): XTX asks deep questions about ML fundamentals. You might be asked to explain the bias-variance tradeoff in detail, describe how gradient boosting works at a mathematical level, discuss the advantages of Bayesian neural networks, or explain when and why you would use causal inference over correlation-based methods. Interviewers will probe your understanding of why algorithms work, not just how to call them from a library. Be prepared to write mathematical derivations on a whiteboard.

Applied Research and Experimental Design: You may be given a problem — "Here's a dataset of market data. How would you build a model to predict short-term price movements?" — and asked to walk through your entire approach. Interviewers evaluate your ability to formulate the problem correctly, select appropriate methods, identify potential pitfalls (data leakage, non-stationarity, transaction costs), and reason about evaluation metrics. This is where practical experience with real data analysis shines.

Systems Design and Coding (Engineers): Engineering interviews focus on building high-performance, reliable systems. You might be asked to design a real-time data pipeline, optimize a critical code path for latency, or architect a distributed system for model inference. Deep C++ knowledge (memory management, concurrency, SIMD) is expected for performance-critical roles. Python proficiency is needed for research infrastructure roles.

Statistics and Probability: Even for engineering roles, XTX expects comfort with statistical reasoning. Researchers face rigorous questions on estimation, hypothesis testing, and stochastic processes. You may be asked to derive estimators, reason about convergence rates, or discuss the statistical properties of time series models.

Research Discussion: Expect to discuss recent ML papers you've read, research directions you find interesting, and how you stay current with the field. XTX wants people who are genuinely passionate about research and can engage in substantive discussions about the state of the art.

XTX's interviews probe ML knowledge at a depth that surprises many candidates. Surface-level familiarity with popular algorithms is insufficient — you need to understand the mathematical foundations, design choices, and failure modes of the methods you use.

Start with "Pattern Recognition and Machine Learning" by Bishop or "Elements of Statistical Learning" by Hastie, Tibshirani, and Friedman for rigorous foundations. Make sure you can derive key results from first principles — backpropagation, the bias-variance decomposition, the EM algorithm, SVD. Practice explaining complex algorithms as if teaching them to a colleague.

Go beyond the basics into areas relevant to trading: time series modeling, online learning, reinforcement learning, and causal inference. Read recent papers from NeurIPS, ICML, and quantitative finance journals. If possible, implement key algorithms from scratch rather than just using library calls — this builds the deep understanding XTX tests for. Having a portfolio of ML projects with clear documentation and results gives you concrete examples to discuss in interviews.

Even for research roles, XTX values the ability to write production-quality code and understand systems constraints. For engineering roles, systems expertise is the primary evaluation criterion.

Strengthen your Python skills by working on data-intensive projects: build efficient data pipelines, optimize numerical code with NumPy vectorization, and learn profiling tools. For C++ roles, study "Effective Modern C++" by Scott Meyers and "C++ Concurrency in Action" by Anthony Williams. Practice building high-performance, concurrent systems and understand how hardware (caches, memory hierarchy, SIMD) affects software performance.

Work on projects that combine ML and systems — for example, building a real-time model serving system, implementing an efficient feature store, or optimizing a training pipeline for speed. These projects demonstrate the intersection of skills that XTX values most. Review XTX compensation data to understand the reward structure and schedule a coaching session for personalized guidance.

XTX interviews include in-depth research discussions where you'll need to explain your past work clearly and engage in substantive technical debates. Strong communication skills differentiate good candidates from great ones.

Practice presenting your research concisely — for any project, be able to explain the problem, approach, results, and limitations in 5 minutes or less. Prepare for probing follow-up questions: Why did you choose this approach over alternatives? What would you do differently? How would you validate this in a non-stationary environment?

Stay current with the ML research community. Read 2-3 papers per week from top venues and practice summarizing them to colleagues. When discussing papers in interviews, demonstrate critical thinking — identify strengths, weaknesses, and potential applications to trading. XTX wants researchers who can evaluate ideas independently, not just implement published methods. Joining reading groups or presenting at seminars builds this muscle effectively.

For researchers targeting XTX's uniquely demanding ML-focused interviews, Quant Blueprint's coaching program provides mentors with deep quantitative research experience at top firms. Our team of 10 quant traders and researchers offer mock research discussions, critique your project presentations, and help you articulate your work at the level XTX expects — bridging the gap between strong ML skills and the polished interview performance that secures offers.