Capital One & Columbia University: AI research awardees 2025

The Columbia Center for AI and Responsible Financial Innovation (CAIRFI) and Capital One are proud to announce the recipients of their 2025 AI research awards. Now in its second year, this collaboration builds on the early impact of the 2024 CAIRFI research fellows and awardees, who helped shape the center’s mission to advance responsible AI and innovation in financial services.

This year, the center issued a Call for Proposals, seeking faculty AI research on the following key areas:

• Synthetic data generation for large language model (LLM) pre-training and post-training
• Multi-agent LLMs for complex reasoning
• Training LLMs for reasoning at inference time
• Reasoning in sequential decision settings
• Multimodality beyond vision and language: LLMs with tabular data and time series
• Generative AI for synthetic financial data
• Techniques to improve generalization/performance from pre-training to downstream use cases

AI research award recipients of the 2025 to 2026 academic year

Following careful consideration, the selected AI research proposals and faculty have been announced as part of Capital One’s AI partnership with Columbia University.

Trustworthy lending: explaining group differences

This project aims to uncover potential causes underlying disparities in loan application outcomes among different demographic groups. Methodology and tools stemming from this research aim to support a better understanding of transparent and suitable lending practices that can benefit both banks and borrowers throughout the industry.

DoubleAgents: using AI agents for task completion and evaluation

DoubleAgents is a general-purpose AI agent system designed to coordinate and allocate human effort in complex, multi-person tasks—such as organizing events, making collective decisions, or managing team-based projects. The system uses AI agents both to complete coordination tasks (e.g., negotiating, emailing, scheduling) and to simulate human behavior for evaluation, allowing it to adapt to strategic, unpredictable and socially nuanced interactions.

By learning from real-world feedback and simulation-based evaluation, DoubleAgents improves coordination outcomes while reducing the burden on human organizers. Beyond AI research, this work contributes to scalable human-AI collaboration tools that can improve productivity, equity and responsiveness in domains like education, civic planning and organizational management.

Reinforcement learning for training LLMs to engineer tabular features

Data scientists can often spend large amounts of time engineering new features out of existing columns in tabular datasets. This procedure often leads to large performance boosts for downstream predictors trained on these features. Automating feature engineering would be valuable, but supervised learning approaches are difficult since ground-truth best feature set for any given tabular dataset may be unknown. This work explores the design of reinforcement learning algorithms for fine-tuning LLMs to perform feature engineering. LLMs have a strong advantage over alternatives because they can read textual column headers and benefit from pretraining. This general approach may be applicable across data science applications broadly and there are tentative plans to test it out in applications ranging from medical diagnosis to finance.

Foundation model for adaptive experimentation

The ability to navigate new environments and make robust decisions is a central hallmark of intelligence. This project advances AI’s societal impact by proposing a new approach to agentic systems that interact with the real world and continuously improve from feedback. We propose to develop techniques for serializing structured interaction data into natural language format, and build tabular foundation models capable of handling long experiences. Our framework provides the groundwork for AI agents that design their own learning curriculum: exploring the environment by quantifying its own uncertainty and taking actions to actively resolve it.

Secure CodeLLM with financial reasoning and compliance

We propose a secure, domain-specialized language model framework designed for critical financial applications, where generic AI systems often fall short. It tackles sector-specific risks such as unauthorized transactions, compliance violations and logic errors that can compromise critical systems in banking, asset management and payments. By grounding its reasoning in executable financial logic and regulatory constraints, the model ensures outputs that are not only syntactically correct but also semantically aligned with institutional policies and legal requirements. A key innovation is its ability to handle both precisely defined rules (e.g., capital adequacy thresholds) and context-dependent principles (e.g., fair lending or risk appetite), blending formal verification with adaptive, judgment-based reasoning. The result is a robust and trustworthy foundation for deploying AI in financial workflows—one that prioritizes correctness, compliance and security by design.

Advancing responsible AI research and talent in financial services

Capital One is committed to advancing responsible AI research and developing top AI talent through strategic academic partnerships. Our collaboration with Columbia University and the CAIRFI initiative reflects our dedication to shaping the future of AI in financial services, supporting cutting-edge research and fostering innovation at the intersection of technology, ethics and finance.

• See how our AI research is advancing the state of the art in AI for financial services.
• Learn how we’re delivering value to millions of customers with proprietary AI solutions.
• Explore AI research jobs and join our world-class team in changing banking for good.