Trading / Financial Systems

This category is intentionally small — one actively developed system, not a portfolio of trading toys. The goal isn’t to list strategies; it’s to document real production trading infrastructure and what actually happened when it ran against live markets.

PhantomArbiter

PhantomArbiter is a real-time Solana DEX arbitrage engine. It aggregates price data from Jupiter, Raydium, Orca, and Meteora over WebSocket, detects cross-DEX, triangle, and statistical arbitrage opportunities, and executes via JITO bundles to avoid MEV front-running.

It runs in three modes:

• Paper: simulate trades without capital
• Monitor: detect opportunities but don’t execute
• Live: execute real trades via JITO bundles

The Honest Result

A live run lost $4 over 400 trades. The strategy detection worked; the execution didn’t. RPC latency and network conditions ate the edge before the bundle landed. That’s a real finding, not a strategy failure — it means the technical foundation is sound, but the economic barriers are practical: dedicated validator access, bundle relay redundancy, and lower-latency infrastructure.

That’s the difference between a backtested idea and a system that actually touched a live market.

Built with Python, TypeScript, Rust, FastAPI, and the Solana SDK. Real money, real lessons, not a simulator.