I build rigorous, inspectable systems for simulation, automation, data quality, and local-first workflows — mostly in Rust, with Julia and Svelte where they fit.
I am drawn to complex problems, purposeful experimentation, and turning ambitious ideas into dependable tools. The common thread is practical innovation with enough transparency and evidence to earn trust.
- FerrousLearn — dependency-free, from-scratch machine-learning algorithms in Rust.
- Surg_sim — Monte Carlo operating-room throughput simulation with pluggable duration estimators.
- surg_front — SvelteKit scheduling interface for the Surg_sim prediction API.
- qaprs — validated CSV-to-LaTeX reporting for laboratory assay data.
- SimulationService — reproducible discrete-event service simulation in Julia.
- Reproducible models and deterministic tests
- Explicit data validation instead of silent coercion
- Focused components with documented interfaces
- Local-first tools that remain understandable and operable
- Useful experimentation grounded in evidence and honest limitations
More projects and infrastructure work: ascentlogic.xyz