Quant Systems — Computational Finance in C++

Level 9 of Baruch's MFE C++ track is the computational-finance module: implement the canonical options-pricing methods in well-structured C++ with STL and Boost. I built it as one continuous codebase across six homework groups (A–F), spanning seven months. Exact pricing and Greeks, perpetual American options, Monte Carlo with standard-error tracking, finite-difference methods, and Excel integration — one library, six entry points.

1. 01

   Exact pricing layer (A & B)

   Implemented the generalized Black-Scholes formula for European calls and puts with cost-of-carry, plus all five primary Greeks (delta, gamma, vega, theta, rho) — both analytical and via finite-difference approximation. Extended to perpetual American options. Wrapped with a matrix pricer that runs parameter sweeps over input vectors. Modular header/source separation, with put-call parity used as a runtime sanity check.

2. 02

   Monte Carlo layer (C & D)

   Built the SDE simulation framework for one-factor European options: discretize [0,T] into N steps, simulate M paths under geometric Brownian motion, discount the average payoff. Then added standard deviation and standard error tracking so the convergence behavior is visible — what NT/NSIM combination is actually needed to hit two decimal places of accuracy on Batch 4 stress data.

3. 03

   FDM + visualization (E & F)

   Plugged in the explicit-Euler finite-difference method as a third pricing strategy and tested against the same batches. Wired Excel output via the provided package so results from any of the three methods could be dropped into a sheet and compared side-by-side. The accuracy / efficiency / stability triangle finally felt visceral instead of abstract.

A unified, six-module C++ codebase that prices European, American, and perpetual options across three numerical methods, with Excel-output and stress-testing instrumentation. Hands-on exposure to STL, Boost, finite-difference stability conditions, and the design patterns institutional pricing libraries actually use. The real lesson: in computational finance the math isn't the hard part — the architecture is. Where to put a virtual function vs a template is the actual quant-engineering skill.