03 / Quant & applied maths

Crucible Trader

Quant backtesting platform with statistical validation

Modular quantitative trading platform: a deterministic backtest engine with a custom strategy DSL, statistical validation (permutation tests, walk-forward analysis), risk guardrails (daily loss caps, position limits, drawdown kill-switches), quant metrics (Sharpe, Sortino, max drawdown, CAGR) and a C++ pricing library, running as an async job worker with Parquet artifact persistence.

Why it matters

The hard part of backtesting is not the engine; it is not fooling yourself. Permutation tests and walk-forward analysis are there to kill overfit strategies before they cost anything.

Built with TypeScript, Next.js, C++20, Python, gRPC, DuckDB, Parquet.

all work