// the find
avhz/RustQuant
Rust library for quantitative finance.
RustQuant is a Rust library for quantitative finance covering option pricing, stochastic processes, fixed income, and basic ML. It targets quant developers who want performance and type safety without reaching for QuantLib or Python. The scope is ambitious — AAD, Heston, Longstaff-Schwartz, term structure models — but it's a solo side project, not a production trading library.
The autodiff (AAD) implementation is the most interesting piece — computing gradients via reverse-mode AD is genuinely useful for Greeks and calibration, and it's rare to see it done cleanly in Rust. The stochastic processes module is broad: standard, arithmetic, fractional, and geometric BM, CIR, OU, Vasicek, Hull-White, SABR, Heston, Merton jump diffusion — all in one place. The workspace is well-structured with each module as its own crate, so you can pull just what you need. Longstaff-Schwartz for American option pricing via least-squares Monte Carlo is a non-trivial inclusion that most hobby quant libs skip.
The ML module is vestigial — linear regression, logistic regression, and k-NN don't belong in a quant finance library and signal scope creep more than genuine utility. No async, no parallelism — Monte Carlo simulations on a single thread will hurt for anything realistically sized. The limit order book in the trading module is essentially a skeleton; calling it production-ready would be generous. It self-describes as a 'free-time project' and the last push was six months ago, which means any bugs you find are yours to fix — the issue tracker is not moving fast.