AI in Systematic Trading

Systematic trading has long depended on technology to identify opportunities and manage risk. From the adoption of rule-based algorithms in the 1980s to today’s advanced machine learning platforms, innovation has consistently reshaped financial markets. Artificial intelligence (AI) now represents the latest—and arguably most transformative—development, enabling hedge funds to analyse unprecedented volumes of data, adapt models dynamically, and execute trades with precision. Firms that have invested heavily in infrastructure and talent are positioned to capture this edge, while others that lack the same technological foundations face increasing disadvantages. The result is a widening gap between funds capable of embedding AI effectively and those constrained by legacy systems or limited expertise. According to a 2023 Market Makers survey of the top 50 Hedge Fund Managers, 9 out of 10 Hedge Fund Traders use AI to achieve portfolio returns.

AI is being integrated across every stage of the systematic investment process:

Signal Generation: uncovering hidden structures in financial and alternative data.

Model Adaptation: refining strategies in real time to reduce signal decay.

Portfolio Diversification: applying varied techniques across multiple asset classes, styles and horizons.

Execution Optimisation: supporting split-second trading decisions and lowering transaction costs.

Beyond classical machine learning, advanced subsets play increasingly important roles. Deep learning enables complex pattern recognition in high-dimensional datasets, while natural language processing (NLP) extracts insights from earnings transcripts, news, and even audio sources. Together, these methods allow systematic managers to integrate broader, less-structured information into trading models.

Opportunities and Risks

The promise of AI in systematic trading lies in uncovering new alpha streams, building more adaptive signals, and improving resilience across market regimes. Yet its risks are equally material:

• Overfitting to historical data may erode predictive power in live markets.
• Contextual blind spots arise when models lack human intuition under unprecedented conditions.
• Homogenisation of strategies can reduce differentiation as similar AI methods proliferate.

Central to both opportunity and risk is data quality. High-quality data must be accurate, timely, and relevant, while poor inputs amplify risk and produce unreliable predictions. This has elevated data governance to a strategic priority, with leading firms dedicating entire teams to sourcing, cleaning, and validating datasets.

Strategic Implications

For systematic managers, AI is no longer a peripheral tool but a structural driver of competitiveness. Advantage increasingly depends on three interdependent factors:

1. Infrastructure: Scalable platforms capable of supporting advanced computation.
2. Data: Reliable, diverse, and continuously validated inputs.
3. Talent: Researchers and engineers able to embed AI responsibly into investment processes.

Firms combining these elements are best positioned to generate sustainable alpha and manage the inherent risks of model-driven trading.

Case Studies:

Case Study 1: Squarepoint Capital – Scaling Machine Learning in Systematic Trading

Squarepoint Capital’s Data Science team demonstrates how machine learning can be industrialised within a global trading framework. Established in 2017, the unit has grown to 14 specialists across New York, London, and Hong Kong, with the Hong Kong hub focused on equities and futures research.

The group develops predictive signals using supervised and deep learning models, applied to traditional and alternative datasets such as supply-chain metrics, internet search volumes, and consumer sentiment scores. Models operate across equities, index and commodity futures, FX, and volatility, forecasting returns over horizons from five minutes to ten days.

Rather than trading directly, the team provides signals to other Squarepoint groups, which can either subscribe to individual alphas or adopt entire AI-enabled strategies. This modular structure allows scalability and portfolio flexibility.

Although reinforcement learning remains under review, the absence of a robust simulation environment has led the group to prioritise scalable ML/DL techniques. Plans to onboard quantitative developers in 2025 suggest future expansion into more complex architectures.

Squarepoint illustrates how systematic managers can institutionalise machine learning by structuring it as a collaborative, globally distributed signal provider, producing differentiated alpha across regions and time zones.

Case Study 2: Man AHL – Organisational Complexity in AI Integration

Man AHL was an early adopter of machine learning, embedding ML-driven strategies into client programs as early as 2014. Yet recent organisational changes reveal the difficulties of sustaining innovation at scale.

In 2023, under CEO Robyn Grew and CTO Gary Collier, Man Group launched a central Data and Machine Learning department led by Tim Mace. Its flagship initiative, ManGPT, is an internal generative AI platform inspired by ChatGPT. Adoption has been notable: 40% of employees use the tool, and around 20% of its suggested code lines have been implemented in production. The long-term vision is an “alpha assistant” capable of analysing datasets, forming hypotheses, and prototyping strategy code.

At the same time, however, AHL’s specialist ML research team was disbanded. Its members were dispersed into broader units such as the Core Strategies Team (trend-following, macro, and risk parity) and a new Fast Trading Team. This redistribution diluted ML focus, frustrated researchers, and drove attrition to firms like Tower Research. Legacy models remain in use, but often without full documentation or iteration, raising oversight and transparency concerns.

The contrast is striking: while ManGPT enhances productivity, the absence of a central ML research hub has hindered direct alpha innovation. AHL’s case underscores the industry-wide tension between building firm-wide AI tools and embedding machine learning as a dedicated research engine. Without the latter, advanced AI risks becoming infrastructure support rather than a true driver of systematic performance.

Conclusions, Reflections & Industry Outlook

The evolution of AI and machine learning in finance has created both transformative opportunities and organisational challenges. This report explored the technical foundations of AI in systematic trading, examined case studies of adoption, and highlighted the widening divide between technology-first proprietary firms and traditional hedge funds.

Structural & Cultural Gaps

Prop firms like Jump Trading and Hudson River Trading integrate researchers, developers, and traders into tightly coordinated teams, enabling rapid iteration and deployment. By contrast, hedge funds often struggle with fragmented structures and slower decision-making, as shown in the Man AHL example. The result is reduced collaboration, diluted expertise, and higher attrition of ML talent.

Technological Leadership

Prop firms also benefit from greater computational power, data accessibility, and low-latency infrastructure, enabling them to test and deploy strategies with unmatched speed. Hedge funds have begun experimenting with in-house platforms like ManGPT, but these remain early-stage compared to the seamless integration achieved by their proprietary peers.

The Path Forward

Despite these challenges, hedge funds have unique advantages — large capital bases, long-term investment horizons, and diversified client mandates. Closing the gap requires:

• Building centralised, well-resourced ML teams with engineering support.
• Developing collaborative workflows across research and trading.
• Investing in data governance and scalable infrastructure to accelerate innovation.

Outlook

The future of systematic trading will be shaped by firms that align people, processes, and technology. AI’s transformative potential is clear, but its impact depends as much on organisational design as on algorithms. Success will belong to those able to innovate at scale, retain top talent, and integrate machine learning seamlessly into their strategic vision.