Our projects - Focused software projects across multiple domains.

Orphos uses an unsupervised learning algorithm to train a statistical model on an input prokaryotic genome sequence and then predicts gene locations based on that model. Because the training is unsupervised, it does not require externally curated models, the algorithm adapts directly to the features of each genome.

The Rust implementation splits work across CPU cores using Rayon for parallel execution, improving throughput on multi-core machines. Even in single-threaded benchmarks, Orphos matches or slightly outperforms traditional C implementations of Prodigal.

Full Human Health works by taking a user’s raw DNA data file—a machine-readable representation of genetic variants—and running it through a variant interpretation engine that associates genotypes with published scientific findings. These associations are filtered and structured into a comprehensive report covering areas such as health risks, and nutrition traits.

PurgeCSS works by analyzing your application’s source files (HTML, JavaScript, templates, or framework-specific files) to determine which CSS selectors are actually used at runtime. It then compares this usage against your stylesheets and removes selectors that are never referenced.

The tool is framework-agnostic and integrates into modern build pipelines, making it suitable for a wide range of frontend stacks. By eliminating unused CSS before deployment, PurgeCSS helps reduce bundle size, improve load times, and keep stylesheets maintainable as applications grow.