bakome-hub/bakome-crypto-quant-engine

This repository advertises itself as a crypto trading engine written in Rust. The README is trilingual, repeating the same description in French, English, and Spanish. The author, who goes by BAKOME, says the project is a single Rust file of around 1,800 lines that you compile directly with the rustc compiler and run as one executable. No external libraries are listed; the README claims the engine has zero external dependencies. The README describes six modules. A multi-exchange collector is meant to pull market data from Binance, Bybit, OKX, and Coinbase. A feature module computes 128 numeric inputs from that data, including returns, volatility, momentum, funding rates, open interest, cumulative volume delta, and liquidation events. On top of those features sits what the README calls a Mixture of Experts with 256 experts, said to specialize in volatility, trend, funding arbitrage, open interest flow, and similar areas. A risk engine is described with a 95 percent Value at Risk calculation, a three percent drawdown limit, and adaptive position sizing. An execution module lists per-exchange latencies in microseconds. Finally, an auto-reinvestment module splits returns 70 percent into trading, 10 percent into GPU, 15 percent into research, and 5 percent into staking, and a CSV logger writes every executed order to disk. The six trading pairs targeted are ETHUSDT, SOLUSDT, BNBUSDT, XRPUSDT, DOGEUSDT, and AVAXUSDT, each routed across the four listed exchanges. Installation is two commands: compile the file with rustc using the -O3 optimization flag, then run the resulting binary. The README ends with promotional material rather than technical detail. There are referral links to two brokers, donation addresses for Bitcoin, Ethereum, Solana, and TRC20 USDT, a Drips sponsor link, and a hardware wish list that mentions a laptop with 16 gigabytes of RAM, a 4G or 5G router, a monitor, and a solar power bank. The author notes that the whole project was built on a Pixel 4a 5G phone without a laptop or fixed Wi-Fi. Readers should treat the README's performance numbers and capability claims with caution. The document is short on verifiable detail: there is no sample backtest, no benchmark data, no test suite link, and no description of how the engine actually places orders against live exchange APIs without any networking dependency. It reads more as a personal showcase project than a battle-tested system.