Earlier this fall, Ramit Sawhney, Global Head of Core AI and ML at Tower Research Capital, spoke at Columbia University’s Mathematics of Finance Fall Practitioner Seminar. His session highlighted the next frontier of AI agents and multi agent systems, and how these rapidly evolving technologies are beginning to reshape quantitative finance.
The Expanding Landscape of AI Agents
Ramit’s talk focused on how large language models are transitioning from passive predictive models into active computational agents that can reason, use tools, and interact with complex environments. Frameworks such as the Model Context Protocol enable models to connect dynamically to services, structured data, and human feedback.
Multi-agent systems and emerging agentic mesh architectures extend this concept further by enabling many specialized agents to collaborate and refine ideas across shared context. This creates a foundation for adaptive, self improving systems that accelerate research and engineering workflows.
Applications Across Trading, Research, and Engineering
Across Tower’s trading, research, and core engineering teams, AI agents and LLM powered systems are experiencing broad and accelerating adoption. These technologies influence how ideas are generated, evaluated, and deployed, whether in strategy research, quantitative analysis, or large scale infrastructure.
AI agents now support a variety of internal workflows by helping teams explore complex datasets, coordinate multi step analysis, and interface more naturally with tools and systems. As these technologies mature, they will lead to more adaptive processes, deeper automation, and faster iteration across the firm.
In the future, multi agent systems and agentic mesh designs are expected to create more fluid collaboration between humans, models, and computational services. These systems will support both exploratory research and mission critical engineering by providing flexible, context aware environments that evolve with the needs of the work.
From Research to Deployment: A Critical Bridge
Ramit noted that the gap between conceptual research and production deployment is shrinking. AI agents are no longer viewed only as experimental tools; they are becoming practical components of real world decision systems.
The next phase of progress involves ensuring robustness, transparency, scalability, and verification. Responsible deployment requires systems that operate reliably under shifting conditions, maintain clear reasoning paths, integrate with risk controls, and meet the performance standards of high throughput environments.
A Shared Future Between Academia and Industry
Ramit’s presentation continued Tower’s long standing collaboration with Columbia and the broader academic and research community. Earlier this year, Tower CEO Albert An participated in the university’s Mathematics of Finance Spring Practitioner Seminar Series. These conversations reflect Tower’s belief that progress in quantitative finance is driven by strong connections between academic theory and real-world application, including the interplay of mathematics, machine intelligence, and large-scale engineering.
The next generation of practitioners will increasingly work with AI driven systems that support research, analysis, and system design. As multi agent approaches mature, coordinated computational agents may assist across the research lifecycle and help teams explore complex problems more efficiently.
Collaboration with academic institutions remains essential to advancing these capabilities and ensuring their practical impact. Tower looks forward to continuing this dialogue and to welcoming talented researchers and engineers who are excited to work at the frontier of AI and quantitative finance.
Looking Ahead
Tower values the opportunity to exchange ideas with the academic community and to help shape the trajectory of AI in quantitative finance. As AI agents, LLM supported research, and multi-agent systems become foundational technologies, Tower will continue to explore and refine how these capabilities can drive innovation across trading, research, and engineering.