Saumik Dana

Saumik Dana

Quantitative Researcher | Production ML Systems | PhD

US Green Card Holder

Summary

Computational scientist with 3 years of experience building production-grade machine learning systems from modeling and validation through deployment and live monitoring. PhD from UT Austin with expertise in systematic trading, cloud infrastructure, and algorithmic optimization.

Experience

Feb 2024 – Dec 2025

Quantitative Researcher

Asset Management Firm – Stamford, CT

Built end-to-end systematic trading infrastructure on AWS with serverless architecture (Lambda, EventBridge, S3, DynamoDB). Deployed automated intraday and weekly options strategies using foundation models, spike detection, and regime-adaptive rebalancing.

Performance: 1.5+ Sharpe, 10% max drawdown, 10% annualized alpha
  • Developed ML pipelines with Chronos embeddings, SHAP feature selection, and Bayesian optimization (TPE)
  • Built RAG-enhanced LLM signal generators from FOMC statements using LangChain and FinBERT sentiment
  • Implemented Hidden Markov Models, genetic algorithms, and multi-objective optimization (NSGA-II) for portfolio construction
  • Optimized data processing with vectorization and Numba JIT, achieving 100-1000x speedups
  • Created real-time React dashboard for volatility surface monitoring with DynamoDB backend
Aug 2023 – Nov 2023

Computational Engineer

VISIE Inc. – Austin, TX

Contributed to robotic surgical navigation platform. Implemented TCP/UDP protocols for robotic arm control and Azure deployment pipelines. Supported successful Series A demonstration.

Aug 2022 – Mar 2023

Computational Lead

Sophelio – Austin, TX

Adapted physics-informed modeling from fusion experiments to financial time series. Built production MLOps pipeline with sparse regression for PDE discovery, Bayesian optimization for signal thresholds, and automated execution.

Technical Skills

ML Tools

  • PyTorch, Scikit-learn, XGBoost
  • HuggingFace, SHAP, Optuna
  • Kneed, Pymoo, d3rlpy, EconML

LLMs & NLP

  • LLM-based inference, RAG
  • LangChain, structured prompting
  • LLM evaluation pipelines

Cloud & MLOps

  • AWS Lambda, S3, DynamoDB
  • EventBridge, CloudFormation
  • Docker, GitHub Actions

Programming & Data

  • Python (Pandas, NumPy, SciPy)
  • JAX, Numba, React
  • Git, Linux/Bash

Specialized ML

  • Bayesian optimization
  • Hidden Markov Models
  • Anomaly detection, RL

Optimization & Algorithms

  • Quadratic programming (CVXPY, SLSQP)
  • Genetic algorithms
  • Pareto fronts, NSGA-II

Education

PhD

Doctor of Philosophy in Engineering Mechanics

University of Texas at Austin, Austin, TX

MS

Master of Engineering in Mechanical Engineering

Indian Institute of Science, Bangalore, India

BS

Bachelor of Engineering in Mechanical Engineering

University of Mumbai, Mumbai, India

Selected Publications

  • Dana, S., Ganis, B., and Wheeler, M.F. (2018). A multiscale fixed stress split iterative scheme for coupled flow and poromechanics in deep subsurface reservoirs. Journal of Computational Physics, 352, 1-22.
  • Dana, S., Zhao, X., and Jha, B. (2022). A two-grid simulation framework for fast monitoring of fault stability and ground deformation in multiphase geomechanics. Journal of Computational Physics, 466, 111405.

Beyond the Numbers

Beer Montage Road Trip Adventures