PROJECTS
Selected works in quantitative finance, engineering, and web development.
Quantitative Finance & Machine Learning
American Option Pricing Under Time-Varying Rough Volatility
Developed a hybrid pricing framework for American options that integrates dynamic Hurst exponent forecasting, regime-switching volatility engines, and signature-kernel approximations. This approach achieves more stable, accurate, and computationally efficient pricing under rough volatility conditions compared to traditional models.
Key Highlights
- MODELING
Designed a multi-regime rough volatility model tailored for early-exercise derivatives.
- ENGINEERING
Engineered an experimental pipeline analyzing pricing accuracy across moneyness, maturities, and volatility regimes.
- BENCHMARKING
Benchmarked the hybrid model against classical finite-difference and rough volatility approaches.
- IMPACT
Explores practical implications for volatility surface shape, early-exercise premiums, and stress scenarios.
E-Commerce & Web Applications
A premium direct-to-consumer e-commerce platform for 'Krypop', a bold, globally-inspired popcorn brand. The site features a complete shopping experience with custom bundles, subscriptions, and a high-conversion checkout flow.
Key Highlights
- BRANDING
Developed the 'Popcorn That Bites Back' identity, focusing on Indian/Asian fusion flavors like Spicy Sweet Fusion and Sichuan Pepper Buzz.
- FEATURES
Full e-commerce functionality including cart management, discount codes, subscription models, and 'Spice Squad' email capture integration.
- GROWTH
Optimized for conversion with social proof, bundle offers, and seamless mobile responsiveness.
Financial Technology & Trading Systems
VertexLadder
A production-grade trading system with ultra-low latency order processing.
Key Highlights
- Key Features
Processes 1.8M+ orders/sec with <1μs latency, FIX 4.4 protocol support, lock-free concurrency with sharded SPSC queues, custom LIFO object pooling
- Performance
~2.5M ops/sec for order add/cancel, P99 latency of 0.67μs
- Infrastructure
Multi-threaded architecture, zero-allocation hot-path, cTrader integration
Educational Platforms
QuantVerse
Comprehensive educational platform democratizing quantitative finance education.
Key Highlights
- Features
Block-by-block lesson viewer with progress tracking, simulation-focused interactive blocks with real-time calculations, Interactive Curriculum Builder (ICB), Plotly.js visualizations, discussion forums, RSS financial news aggregation
- Architecture
Role-based access control (Student/Instructor/Admin), WebSocket for real-time updates, multi-layer caching strategy, Pyodide for client-side Python execution
- Deployment
Vercel with CI/CD
PEWP
Key Highlights
- INSPIRATION
The contrast between Earth's steady rhythms and Mars' extreme fluctuations sparked the idea: if we can learn the language of weather from these two, can we translate it to worlds light-years away?
- HOW WE BUILT IT
We combined NASA Exoplanet Archive, Mars InSight, and Copernicus Earth data. After cleaning and resampling to create consistent datasets, we trained a Random Forest Regressor to predict atmospheric conditions.
- ACCOMPLISHMENTS
Built a complete pipeline with high accuracy (R² = 0.9421). Successfully demonstrated that Earth and Mars data is sufficient to make informed forecasts about exoplanet atmospheres.
PEWP is a machine learning framework that forecasts temperature, pressure, and wind on exoplanets. It uses atmospheric data from Earth and Mars, then extrapolates to planets where only a few constants are known. By training on real, detailed data, the model can make informed predictions about environments far outside our reach.
Lawrenceville Student Transcript Analysis Tool
Academic performance analysis tool for parsing transcripts and checking graduation requirements.
Key Highlights
- Features
PDF transcript parsing, grade extraction, graduation requirement checking, dark/light mode, multi-student processing, CSV export
- Deployment
Live at lawrenceville.netlify.app