Eastdown Park places ML engineers at systematic hedge funds and prop trading firms — exclusively. Retained search, deep network, no generalists.
The ML engineers suited to a systematic trading environment represent a distinct and genuinely small subset of the broader ML talent pool. They combine research-grade modelling capability with an understanding of financial data — tick data, order book dynamics, alternative data signals — and the engineering discipline to deploy models into production systems where latency, reliability, and P&L accountability are non-negotiable. Across London, New York, and Amsterdam, there are perhaps 300–500 people who meet that bar. They are overwhelmingly passive.
Most ML engineers can build and train models. Far fewer can take a predictive model through the full pipeline — feature engineering on high-frequency financial data, backtesting with realistic slippage and capacity constraints, deployment into a low-latency production environment, and ongoing monitoring for signal decay. That end-to-end capability is what systematic trading firms actually need, and it is rare.
A candidate who has spent years working on transformer architectures for alpha research or reinforcement learning for execution optimisation knows immediately whether the person approaching them understands what that work involves. Outreach that conflates "quant" with "ML engineer" or treats JAX and PyTorch as interchangeable signals is dismissed before the second sentence. Credibility in the approach is not optional.
Total guaranteed compensation for a senior ML engineer at a top-tier systematic fund — base, guaranteed bonus, and signing payment — regularly reaches £250,000–£320,000 in London and $400,000–$500,000 in New York. Structuring an offer that competes with hyperscaler RSUs, academic freedom, and the brand pull of DeepMind or OpenAI requires a recruiter who understands what these candidates value and how to frame the opportunity accordingly.
When three recruiters approach the same passive candidate about the same role at the same firm, it signals a lack of seriousness and process discipline. In a community where the ML engineering talent pool across systematic trading is genuinely small — where a researcher at Man AHL and a model deployment engineer at XTX Markets almost certainly know each other — that reputational damage is lasting and disproportionate to the short-term convenience of running a wide search.
Every firm a major search firm has placed into becomes off-limits to every other client on their books. At scale, this quietly closes off a significant portion of the talent pool before a search even begins. As a specialist boutique, Eastdown Park carries a deliberately minimal off-limits footprint — which means more of the people you actually want to hire are available to approach on your behalf, without restriction.
Signal development and alpha research infrastructure. Transformer architectures for cross-sectional prediction, NLP pipelines for alternative data, feature engineering on tick and order book data. PyTorch, JAX, GPU cluster management.
Statistical and ML model development applied directly to trading. Analysing large-scale datasets to generate investable signals, backtesting systematic strategies, forecasting price action and volatility. Masters or PhD typical — Mathematics, Physics, Statistics, or CS. Python, C++, R.
Applied research at the intersection of machine learning and quantitative trading. Novel model architectures, reinforcement learning for execution, deep learning for market microstructure. Strong mathematical foundations essential — typically PhD-level from top programmes in CS, statistics, or physics.
Low-level systems programming and ML inference optimisation. Profiling and accelerating model serving pipelines, kernel-level optimisation, GPU/CUDA performance engineering. The specialism at the intersection of ML and microsecond-scale systems.
Building the infrastructure that makes research teams faster. High-performance research platforms, data pipelines for financial and alternative data, tooling for model development and backtesting. The role XTX Markets calls Software Developer, Research Technology.
Research-to-production translation. Backtesting framework development, capacity and slippage modelling, signal decay analysis. Python research stack with C++/Rust production layer. Rare combination of quant rigour and engineering depth.
Architecture leadership across the research and production stack. Team building in a constrained talent market, managing the research-to-deployment lifecycle, and setting technical direction for ML capability within a live trading environment.
Data infrastructure, feature stores, and experimentation platforms. Scalable pipelines for alternative data ingestion, distributed training infrastructure, model versioning and monitoring. The platform that makes the research team faster.
We work with multi-strategy hedge funds, systematic macro funds, quant prop trading shops, and systematic trading firms across London, New York, Amsterdam, and Europe.
We begin every search by mapping the full addressable talent pool — identifying ML engineers currently working in systematic trading environments, alongside transitional candidates from adjacent environments. This mapping is shared with you at the outset.
Every candidate is screened on technical depth, understanding of financial data and market microstructure, and cultural fit for a performance-driven environment. We do not present CVs — we present fully briefed candidate profiles with our assessment attached.
We manage the entire candidate experience from first approach to offer acceptance. In a community this small, how candidates are treated reflects directly on your firm. We protect your brand at every stage.
Throughout the search we provide ongoing intelligence on competing offers, compensation benchmarks, and candidate sentiment — giving you the information needed to move decisively when the right person is identified.
The conversations that matter in this niche — with a researcher deploying transformer-based cross-sectional models, or an infrastructure engineer rebuilding a low-latency inference stack in Rust — require genuine technical understanding. Not surface familiarity. The candidates you want to hire can tell the difference immediately.
We understand the difference between a predictive model and a tradeable signal. Capacity constraints, turnover costs, Sharpe decay, and the challenges of overfitting on financial time series are part of the vocabulary of every candidate conversation we have. We know why a researcher might choose JAX over PyTorch for a particular workload, and what it means to work at the intersection of alpha research and production deployment.
Deploying a model to production in a systematic trading context is not the same as deploying to a web service. Latency budgets are measured in microseconds. Feature computation must be deterministic and auditable. Model drift has a direct P&L consequence. We understand this operating environment and communicate it accurately to candidates — which is why candidates from production-grade ML backgrounds take the conversation seriously.
The integration of alternative data — earnings call transcripts, satellite imagery, web-scraped pricing signals, social sentiment — into systematic strategies has created a distinct sub-speciality: ML engineers who can design and maintain NLP and multimodal pipelines at production scale. We map this talent pool separately, understanding both the technical requirements and the organisational environments where it sits.
Reinforcement learning for execution optimisation — minimising market impact, learning adaptive order routing — is one of the fastest-growing ML specialisms in systematic trading. The engineers working in this space sit at the intersection of RL research and microsecond-scale engineering. We know who they are, where they work, and how to reach them.
Our fee is 25% of total guaranteed first-year compensation — base salary, guaranteed bonus, and any signing payment confirmed at offer. Structured across three milestone tranches.
Our fee is 25% of total guaranteed first-year compensation — base salary, guaranteed bonus, and any signing payment confirmed at the point of offer. Discretionary bonus, equity, and co-investment rights are excluded. Fee details and tranche amounts are discussed and agreed at the point of engagement.
Eastdown Park was built on 16 years of placing buy-side systematic trading technologists — trading system developers, quantitative developers, ML engineers, and researchers — at the firms that define this market. The relationships that matter in this community were built over a long time, not assembled from a database.
16 years of specialist experience placing buy-side systematic trading technologists across London, New York, and Amsterdam — trading system developers, quantitative developers, systematic researchers, and the ML engineering talent that now sits at the centre of how the best firms generate alpha.
Track record of senior placements at some of the most prestigious systematic hedge funds and prop trading firms in the world — including multi-strategy platforms, systematic macro funds, and leading quant prop shops across London and New York. Roles spanning the full technology stack: from signal research and model development through to execution infrastructure and platform engineering.
Deep familiarity with the technical environments candidates come from and move into — PyTorch and JAX research stacks, C++ and Rust production systems, low-latency inference pipelines, alternative data infrastructure, and the specific demands of deploying predictive models in live trading contexts.
Retained and exclusive mandates only. A small number of active searches at any time — which means every search receives direct, senior-level attention from first approach to offer acceptance.
Selected placements from prior firm. Full track record and references available on request.
Initial market mapping conversation: complimentary.
tariq@eastdownpark.com