AI · Quantitative finance · Applied mathematics

Research

I study model assumptions, reasoning quality, and changing financial risk, then build methods and tools to test them.

My current work connects mathematical reasoning in AI, evaluation-data quality, and portfolio risk across market regimes.

Current work

Questions, methods, and contributions.

AI reasoning & evaluation

Generative AI research for Meta · Contract

Active research · Annotation repair is early-stage

Reasoning quality and the data used to judge it

How can we distinguish a model’s reasoning error from a flaw in the task, rubric, or reference solution?

My contribution: I evaluate mathematical and STEM reasoning, partner with researchers on research questions and vendor strategy, and lead early-stage work on annotated-dataset quality within a four-person leadership group advising three vendors.

Methods & approach

Proof checking, prompt and rubric review, reference-solution validation, and analysis of annotation workflows.

Questions I’m investigating

Distinguishing an incorrect argument from a valid alternative solution.
Diagnosing and repairing annotation issues while preserving task intent and identifying when human review is needed.

Explore my AI research

Quantitative finance

Regime Alpha · Independent research & development

Ongoing research · Public mechanics demonstration

Portfolio risk across changing market regimes

When market conditions change, which diversification assumptions weaken—and how does the portfolio’s risk change with them?

My contribution: I develop the quantitative models, Python data pipelines, and analytical workflows connecting portfolio holdings to market factors and regime-dependent risk diagnostics.

Methods & approach

Gaussian and Bayesian mixture models, hidden Markov models, stochastic state transitions, and portfolio factor decomposition.

Questions I’m investigating

How factor relationships and risk contributions differ across inferred market states.
What sample support and modeling assumptions are needed to interpret those differences.

Explore the research and public example

Professional foundation

Testing models against economic reasoning.

My model-risk work at JPMorgan shaped this approach: test statistical robustness, challenge hidden assumptions, and ask whether econometric relationships have an economic justification. I carry that discipline into AI evaluation and independent model development.

Read about my model-risk work

Academic background

Earlier work on distributional modeling and sequential decisions, through optimal transport and optimal control.

Master’s thesis · 2022

NYU Courant · M.S. Mathematics

Optimal transport for financial time series

Undergraduate thesis · 2018

Farmingdale State College · B.S. Applied Mathematics

Optimal control in retirement portfolios

Discuss a research question View resume