A proprietary, institutional-grade high-frequency trading (HFT) desktop application built with Rust and accelerated with WGPU compute shaders. Fully audited, optimized, and ready for deployment.
Titan HFT is not a simple trading script. It is a robust, modular desktop suite engineered for memory safety, concurrency, and mathematical precision.
Utilizes raw WGSL compute shaders to run 150,000 concurrent genetic agent simulations directly on the GPU (Vulkan, Metal, or DirectX12). Pre-allocated memory buffers eliminate dynamic allocation overhead, allowing real-time strategy consensus under high tick volumes.
Generates predictive direction signals by combining advanced mathematical structures: Kalman filtering (Joseph covariance update form), Shannon entropy (market uncertainty), Hurst exponent (fractal memory), Lyapunov exponent (chaos detection), Wavelets, VPIN, and Bayesian Online Change Point Detection.
Features a lock-free, thread-isolated risk auditor running every 1.5s. It implements a 3-level circuit breaker (Caution, Halt, Panic), dynamic stop-losses using Average True Range (ATR), and spreads/Order Flow Imbalance (OFI) gates to prevent bad entries during liquidity gaps.
Employs a background Rayon thread pool running genetic optimization cycles against tick-by-tick historical datasets. It evaluates fitness using out-of-sample constraints and dynamically swaps winning parameter matrices into the live trading thread without restart.
The application features a premium dark-themed Tauri desktop interface integrated with low-overhead WebSockets and local REST APIs for execution tracking.
We provide full transparency. Download the audit reports, connection logs, and backtest results prepared for due diligence.
A complete quantitative and qualitative review of the 48,659 lines of Rust/WGSL code, codebase complexity metrics, modular breakdowns, and IP valuation models.
A clean console log demonstrating the application successfully establishing secure WebSocket connections, initializing GPU WGSL shaders, and parsing Alpaca data feeds.
A detailed 1.3-year historical backtest report showcasing profit metrics, win rate, maximum drawdown, Sharpe Ratio, and execution counts on QBTS tick data.
The winning bidder on Flippa will receive the complete, sanitized source code package with no trailing API keys or private databases. Everything you need to build, test, and host the software is included in the package.
Contains the fully compileable codebase (Rust 1.75+ & Tauri v1), build scripts (`clean_and_package.sh`, `build.sh`), and markdown manuals.
Bidding is now open on Flippa. The sale is conducted via a secure escrow system. You can bid directly or contact us for private negotiations.