This repository contains my solutions to the Goldman Sachs India Hackathon 2025 - Quant, where participants tackled real-world financial challenges involving portfolio hedging, adaptive market-making strategies, and exotic option pricing.
I secured Rank 8 overall.
Objective: Hedge an unhedged portfolio using a set of equity stocks.
Goal: Minimize Value at Risk (VaR) and hedging cost.
- Approach: Utilized LassoCV regression with customized enhancements to determine optimal hedging weights.
- This was the most approachable problem — many contestants achieved high scores during the contest.
- I was ranked 1st in this problem after post-contest evaluation, with a final score of 94.12.
Objective: Build an adaptive quoting strategy using order book data, recent trades, and inventory levels.
- Inspired by the Avellaneda & Stoikov market-making framework.
- Method: Developed a dynamic market maker that adjusts bid/ask spreads in response to market volatility, order imbalance, and current inventory.
- Challenge: This was the most complex and demanding problem, requiring robust modeling and real-time decision-making.
Objective: Price exotic European up-and-out basket options on three correlated assets using Monte Carlo simulation.
Bonus: Calibrate local volatility surfaces using market data from vanilla call options.
- Constructed local volatility surfaces from vanilla option prices.
- Simulated correlated asset paths and implemented efficient Monte Carlo pricing with barrier condition checks.