Research &
Development.
Derivatives pricing, risk modelling, factor replication, and live market tools — spanning stochastic calculus, numerical methods, and empirical asset pricing.
Volatility Surface Explorer
An interactive 3D implied‑volatility surface visualizer for any US‑listed equity. A custom FastAPI backend fetches live options chains from Yahoo Finance via curl_cffi browser impersonation, cleans the data (drops NaN IVs, zero‑OI noise, clamps illiquid outliers), and caches it in memory for 10 minutes. The frontend renders a smoothed Plotly mesh with four Z‑axis metrics (IV, total variance, volume, open interest), four camera presets, 10 color palettes, and side‑by‑side ATM term structure and skew‑slice mini‑charts.
IAQF 2025 — Geeks for Greeks
Our UCLA MFE team placed Top 4 out of 30 teams in the 2025 International Association for Quantitative Finance student competition. We modelled multi‑asset volatility dynamics using GARCH‑DCC and detected bubble regimes with LPPLS critical‑time estimation. The paper covers regime identification, correlation break‑down during stress episodes, and a framework for dynamic hedging under non‑stationary vol.
Geeks for Greeks — UCLA Anderson MFE
Full competition report documenting our GARCH‑DCC + LPPLS bubble detection methodology, empirical results across equity indices, and the dynamic hedging framework that secured our Top 4 finish.
Read the Paper (PDF)Risk Analytics Dashboard
A real‑time portfolio risk monitoring dashboard with interactive VaR/CVaR gauges, GARCH‑based volatility forecasting, factor exposure decomposition across Fama‑French dimensions, cumulative P&L attribution, and configurable stress‑test scenarios. The FastAPI backend streams live market data through WebSocket connections while D3.js renders responsive, theme‑aware visualizations. Designed for institutional risk management workflows with multi‑portfolio support and automated breach alerting.
Commodities Dashboard
A real‑time commodity futures analytics dashboard featuring term structure curves, realized and implied volatility analysis, seasonality patterns, correlation matrices, spread analysis, and position calculators. The FastAPI backend fetches live OHLCV and options data for crude oil, gold, natural gas, copper, silver, and wheat. The React frontend renders interactive Plotly and Recharts visualizations with dark/light theme support and Gemini API integration for enhanced analytics.
Financial Risk Management
Eight projects covering the full risk management pipeline: Component VaR/CVaR decomposition, VaR backtesting through the 2008 crisis (Kupiec test), GARCH vs EWMA vs historical simulation comparison, Stulz multi‑asset exotic‑option Greeks, Basel III compliance deep‑dives on Bank of America, CDS hazard‑rate bootstrapping with piecewise constant lambdas, the LTCM collapse case study, and the Merton structural credit model (distance‑to‑default, recovery rate estimation).
Quantitative Asset Management
Five‑project factor replication suite built on WRDS CRSP/Compustat data. Constructed market excess return (MKT‑RF), risk‑free rate, RMW (robust‑minus‑weak), CMA (conservative‑minus‑aggressive), size (SMB), value (HML), and momentum (WML) factors. Achieved >0.99 correlation with published Kenneth French benchmarks. Final project: a Betting Against Correlation (BAC) factor isolating the correlation component of beta, demonstrating orthogonal alpha over 2010‑2024.
Advanced Computational Methods
Three projects covering the full spectrum of numerical option pricing: binomial and trinomial trees for American options with discrete dividends, variance‑reduced Monte Carlo (control variates, antithetic variables) for Asian and lookback options, implicit/explicit/Crank‑Nicolson PDE solvers for American puts, Vasicek and CIR term‑structure models, and quadrature‑based pricing. Also includes exotic barrier options with rebates and Richardson extrapolation.
FinBERT Sentiment Pipeline for Fed & Macro Communications
Capstone Applied Finance Project for DoubleLine Capital (Sep – Dec 2025). Designed an NLP sentiment pipeline using FinBERT to extract systematic, machine‑readable signals from Fed minutes, FOMC statements, and broader macro communications — generating quantitative features for credit and rates research. Built a Python ETL workflow that ingests, cleans, and tokenizes the unstructured text feed, scores each document on hawkish ↔ dovish axes, and packages the model outputs into reproducible datasets handed off to DoubleLine's quant research team.
Exotic Derivatives & Fixed Income Valuation
Capstone project implementing jump‑diffusion default option pricing (Poisson jumps), down‑and‑out puts under Heston‑type stochastic volatility, CIR and G2++ term‑structure models for bond and swaption pricing, and mortgage‑backed security valuation with the Numerix prepayment model (IO/PO tranches, OAS computation). All solvers built from scratch using full‑truncation Monte Carlo and implicit finite‑difference methods.