Skip to main content

AI Research Lead @ J.P. Morgan · NYC

Founding member of J.P. Morgan AI Research. I build AI systems that learn how financial markets work – from foundation models trained on trade-level data to reinforcement learning for portfolio optimization.

Srijan Sood

Career

J.P. Morgan
AI Research Lead, Vice President · 2023 – Present

Built TradeFM, a foundation model for financial markets, trained on billions of trades, that learns price-formation dynamics across thousands of equities. Deep RL for portfolio optimization, presented across institutional clients; knowledge graphs and graph ML for financial analysis.

AI Research Scientist, Senior Associate · 2019 – 2022

Fairness-aware ML for consumer lending (>17k models in production). Time series forecasting for equities.

Georgia Tech
M.Sc. Computer Science · 2018 – 2019

RL for language model alignment (Dr. Mark Riedl) and RL for trading (Dr. Tucker Balch).

Aerial Intelligence
Machine Learning Engineer · 2017 – 2018

Satellite imagery ML for land-cover classification and crop yield prediction.

Georgia Tech
B.Sc. Computer Science, Highest Honors · 2013 – 2017

Thesis: Domain Adaptation in RL (Dr. Charles Isbell).

Nervana Systems
Machine Learning Intern · 2016

High-throughput ML infrastructure (~70× speedup); neural style transfer.

PayPal
Software Engineering Intern · 2015

Risk data science and analytics.

Software Engineering Intern · 2014

Embedded systems firmware.


Selected Publications

Ensemble Methods for Sequence Classification with Hidden Markov Models
S. Sood*, M. Kawawa-Beaudan*, S. Palande, G. Mani, T. Balch, M. Veloso · arXiv 2024
Ensemble HMMs paper figure
AI in Investment Analysis: LLMs for Equity Stock Ratings
K. Papasotiriou, S. Sood, S. Reynolds, T. Balch · ICAIF 2024
LLM Ratings paper figure
Deep Reinforcement Learning for Optimal Portfolio Allocation
S. Sood, K. Papasotiriou, M. Vaiciulis, T. Balch · ICAPS FinPlan 2023
DRL Portfolio paper figure