Projects

Applied and computational work spanning numerical methods, scientific computing, data engineering, causal inference, and production machine learning.

• Numerical methods

  Regional dynamic input–output model

  JAX solver for a high-dimensional regional dynamic economy with input-output relationships.

  • JAX
  • Python
  • In development
• Scientific computing

  DSGE perturbation solver (Python)

  Built on a SymPy backend, validated on a two-country IRBC model.

  • Python
  • SymPy
  • Code coming soon
• Data engineering

  Interwar trade price-index pipeline

  Laspeyres / Paasche / Fisher indices plus state deflators for U.S. trade, 1927–1935.

  • Python
  • pandas
  • Index theory
  • Code coming soon
• Production ML

  Credit risk models at American Express

  XGBoost models on large-scale transaction data for credit risk, built on PySpark, Hive, and SQL.

  • XGBoost
  • PySpark
  • HiveQL (SQL)
  • Proprietary