BruceBrasseur
I build quantitative trading systems, simulation engines, and research infrastructure for financial markets.
Selected Projects
Trading systems, financial models, and research tools built for quantitative analysis.
Rust Backtester
Backtesting platform with a deterministic Rust engine and a Next.js frontend. Supports MA crossover, z-score mean reversion, Donchian breakout, and pairs strategies on equities, ETFs, and crypto. Signals are computed on bar close, orders fill at the next bar's open, and fees, slippage, and borrow costs are all configurable.
Market Impact Sim
A limit order book simulation built around a price-time priority matching engine. Noise traders and value traders establish a baseline price path, then a large institutional buyer is introduced to measure price displacement. Sweeping across order sizes reproduces the square-root market impact law observed in real markets.
HopScout
Queries DEX factory contracts to discover pools, builds a token graph, and searches for profitable multi-hop swap cycles using live on-chain reserves. Each cycle is simulated locally with the constant-product formula to account for slippage and fees. A ternary search finds the input size that maximizes profit.
BTC GAN Anomaly Detection
Trains a GAN on hourly BTC/USD OHLC data to learn what "normal" price windows look like. A dense autoencoder compresses each timestep into a latent vector, and an LSTM discriminator scores rolling windows. Windows that score below a quantile threshold get flagged as anomaly events.
EvoLoss
A CLI tool that uses genetic programming to discover and evolve novel loss functions for deep learning models. Explores the loss function design space automatically, finding differentiable functions that outperform hand-crafted alternatives.
About Me
I hold a degree in computer science and have been programming for over seven years. My work focuses on quantitative finance and simulation—building backtesting engines, market microstructure models, and anomaly detection pipelines.
I enjoy decomposing complex financial problems into clean, performant code. Whether it’s optimizing a matching engine in Rust or prototyping a trading signal in Python, I care deeply about correctness and efficiency.
I take pride in being able to move between deep quantitative work and practical systems engineering without losing sight of either.