How Agentic AI Is Changing Quantitative Trading in New York and Beyond

Artificial intelligence is changing how quantitative trading teams develop strategies, analyse markets and approach trade execution. But as AI agents take on more technical work, what happens to the expertise that gives trading firms their competitive advantage?

For hedge funds, proprietary trading firms and institutional technology providers, agentic AI presents opportunities to accelerate research and simplify complex workflows. It also raises questions about human oversight, trading infrastructure and the skills required to operate in increasingly automated financial markets.

In this episode of FinTech Focus TV, Toby Babb visits Talos in New York City to speak with Eliad Hoch, Head of Quant Execution Services. Drawing on his experience in systematic trading and digital assets, Eliad explains how his team is adopting agentic AI and shares his vision for the future of institutional trading technology.

Their conversation also explores cryptocurrency market structure, event contracts, 24/7 trading and the opportunities AI could create for the next generation of quantitative finance professionals.

How is agentic AI changing quantitative trading?

Agentic AI allows users to communicate objectives to systems capable of carrying out multiple connected tasks. In quantitative finance, this could change how researchers develop models, test strategies and build trading technology.

At Talos, Eliad describes a significant shift in his quantitative team's working practices. Rather than writing every line of code directly, the team increasingly uses AI agents to support development, concentrating on the intended outcome and reviewing the results.

This is more than a productivity improvement. It changes where quantitative expertise is applied.

Researchers still need to understand market behaviour, identify weaknesses in their analysis and determine whether a proposed solution is commercially useful. AI can accelerate the technical process, but faster research does not automatically produce better trading decisions.

That distinction is also explored by Harrington Starr's Oli Knight in When Everyone Has AI, Where Does the Edge Come From?. Oli argues that as AI makes research and coding more accessible, specialist market knowledge and the ability to recognise worthwhile opportunities become increasingly important.

For quantitative trading firms, the competitive question is therefore shifting from how quickly teams can build technology to how effectively they use it.

Could AI agents transform institutional trade execution?

One of Eliad's most interesting ideas concerns the future of AI-powered trading execution.

He describes a potential workflow in which an institutional trader could instruct an AI agent to execute a substantial cryptocurrency order across selected exchanges within a specified timeframe.

Instead of manually configuring every parameter, the trader would communicate their objectives. The agent could analyse market conditions, assess available execution algorithms and present alternative strategies for review.

The trader would then approve the preferred approach before execution.

Importantly, Eliad presents this as a future direction for Talos, not as a fully deployed autonomous trading capability.

The distinction matters because institutional execution involves considerable complexity. Liquidity, market impact, transaction costs and timing can all influence the outcome of a trade.

AI agents could make these considerations easier to evaluate, but the underlying systems must remain reliable and subject to appropriate controls.

For companies developing trading and exchange infrastructure, this creates opportunities to rethink how professionals interact with execution technology.

Rather than navigating numerous interfaces and configuring individual instructions, traders could increasingly communicate their intentions while AI supports the technical decision-making process.

Why quantitative execution matters in digital asset markets

Eliad's experience in systematic trading provides an important foundation for his perspective.

He distinguishes between identifying a profitable trading opportunity and executing it effectively. 

A promising quantitative strategy can lose much of its value if trading costs, slippage or market impact are not properly managed. Harrington Starr's What Will the Future Hold for Quant Hiring and Retention? report reinforces this point, highlighting how sophisticated execution algorithms, proprietary data and trading infrastructure can give quantitative firms a competitive advantage.

This is particularly relevant in cryptocurrency markets, where liquidity is distributed across different venues.

At Talos, Eliad's team works across pre-trade analytics, execution algorithms and post-trade analysis, helping institutional participants assess and improve their trading activity.

He also challenges the idea that Bitcoin's price provides a complete picture of the digital asset industry.

Tokenisation, real-world assets and new trading instruments are expanding the range of opportunities and infrastructure requirements within cryptocurrency markets.

These developments demand more sophisticated connectivity, market data and execution capabilities.

Mark Price examines this evolution in Crypto Infrastructure in 2026: A Look at Where Things Are Heading, originally published in The Financial Technologist: Influence List 2026. His analysis considers how cryptocurrency venues are adapting their infrastructure to meet institutional expectations around performance, connectivity and reliability.

The connection to Eliad's argument is clear: as digital asset markets develop, execution technology becomes increasingly important to institutional participation.

What opportunities are emerging beyond traditional crypto trading?

Eliad also discusses the growing institutional interest in event contracts and prediction markets.

These instruments allow participants to trade around specific outcomes, potentially creating opportunities for hedging and systematic strategies.

For quantitative traders, differences in pricing between related contracts or venues may present arbitrage opportunities. However, identifying those differences is only the beginning.

Firms also need reliable order book data, suitable execution algorithms and effective risk management to determine whether an apparent opportunity is genuinely tradable.

The conversation illustrates how new markets can create demand for established quantitative techniques while introducing unfamiliar operational challenges.

Another development is the movement towards 24/7 trading and settlement.

Eliad questions why capital should remain constrained by conventional market hours when digital infrastructure could support more continuous activity.

However, he acknowledges that regulation, settlement arrangements and existing financial systems create obstacles.

For institutional markets, continuous trading requires more than extending opening hours. It demands infrastructure capable of supporting execution, settlement and operational controls across the entire trading lifecycle.

Why New York matters for the future of quantitative trading

Recorded at Talos's New York headquarters, the conversation also reflects the city's position within the US quantitative trading ecosystem.

New York brings together investment banks, hedge funds, asset managers, trading technology providers and digital asset businesses. This concentration creates an important environment for developing and implementing institutional financial technology.

Agentic AI adds another dimension to the technical capabilities these organisations require.

Quantitative researchers, trading technologists and software engineers may increasingly need to combine financial market expertise with the ability to develop, supervise and evaluate AI-assisted systems.

These requirements extend beyond New York. Other established US financial centres have their own concentrations of trading businesses and specialist expertise, as reflected in Harrington Starr's wider US FinTech talent hotspots.

For businesses building quantitative trading capabilities, location remains relevant, but the more important question is how effectively they can bring together market knowledge, engineering expertise and emerging AI technologies.

What skills will the next generation of quant traders need?

Eliad offers an encouraging perspective on the impact of AI on early-career quantitative professionals.

Rather than advising young researchers to avoid AI, he encourages them to use it to build projects, test ideas and demonstrate their abilities.

"Do you have an interesting idea? Go and vibe code it."

His advice includes publishing work on GitHub and being able to explain what has been built.

This distinction between producing code and understanding it is particularly relevant as AI tools become more capable.

A candidate who can demonstrate how a trading model works, identify its limitations and explain its practical application offers something beyond the ability to generate technical output.

The same principle connects with Oli Knight's perspective on the US quantitative finance market.

In Harrington Starr's 2026/27 Salary & Hiring Report, Oli discusses the importance of proven quantitative expertise, commercial understanding and capabilities that contribute to trading performance.

Although his commentary addresses the wider US market rather than Talos specifically, it reinforces the importance of demonstrable skills and sound judgement.

As AI lowers some technical barriers, the ability to identify meaningful opportunities, assess risk and apply research to real financial markets could become an even more important differentiator.

For employers, this also raises questions about how quantitative talent is assessed and developed. 

research quality and evidence of independent thinking may provide valuable insight alongside conventional qualifications and experience.

What comes next for agentic AI in trading?

Eliad's longer-term vision involves AI agents supporting increasingly complex trading and operational workflows, with human professionals supervising their activity.

He also highlights the importance of proprietary data and the information available within trading platforms.

This suggests that competitive advantage may depend not simply on access to powerful AI models, but on the quality of the data, execution technology and market expertise supporting them.

For financial institutions in New York and across the US, successful adoption will require careful decisions about technology, governance and specialist capabilities.

The future of quantitative trading is unlikely to be defined by automation alone. It will depend on how effectively firms combine AI with human judgement and reliable market infrastructure.

For businesses developing these capabilities, Harrington Starr's New York financial technology team provides a local connection to the specialist US talent market.