I enjoy writing correct, fast and simple software using Rust ๐ฆ on NixOS.
Interested in predictable low latency trading systems, optimizing for CPU caches, branch predictors and proper atomics synchronization primitives.
Using NixOS to manage my homelab (6 servers, ~400 threads, ~1.7TB RAM) in a declarative and reproducible fashion, running various services, miners and ML-driven tick-by-tick backtests. Gotta pump that load average up. Also more than half of my contributions are directed towards my self-hosted forgejo instance, not showing up on Github.
My Open-Source Repositories Crates on lib.rs:
- ๐ฐ TradeAggregation: Low-Latency trade aggregation crate for user-defined
Candles based on information driven rules. - ๐ฑ LFEST High performance perpetual futures exchange for simulated trading
- โ๏ธ Chainable Tree-like Sliding Features. Basically sequential signal processing including "technical indicators"
- ๐น EhlersIndicators: John Ehlers technical analysis indicators / Filters
- ๐ธ Evolutionary Neural Network implementation called CoSyNE.
- ๐ช gym-rs a pure Rust implementation of OpenAi's Gym environments.
- ๐ TimeSeriesGenerator for synthetic data generation.
- ๐ฐ reservoir_computer Reservoir computers on the CPU
- ๐ whittaker_smoother A perfect smoother