I ship privacy-first software and measurable AI systems.

Work

I build two kinds of things: (1) products people use, (2) AI systems that must be measurable and reliable.

End-to-end product build (UI, backend, cloud).
Privacy-first mindset applied from day one.
Built for real clinic workflows.

What you can expect from a similar build

A clean, minimal UI for one workflow.
Stable APIs, predictable data model.
Deployment and basic monitoring.

I can't share client internals publicly, but these are the system types I routinely build:

Predictive modeling pipelines (training + evaluation + monthly/recurring runs).
Time-series forecasting and feature engineering for seasonality/holidays/promotions-style signals.
LLM workflows: dataset construction, deduplication/quality filters, ablations, fine-tuning, and evaluation gates.

Example: Predictive modeling for regulated industry

Problem: Need monthly forecasts with explainability requirements and audit trail.

Approach: Baseline models first, then feature engineering pipeline, evaluation harness with business metrics, model versioning.

Deliverables: Reproducible training pipeline, evaluation report, deployment plan, monitoring setup, documentation.

Research/portfolio sites, fast, clean, SEO + analytics, deployed.

Research/portfolio site: publications, timeline, contact form, responsive design.
Founder/consultant site: services, case studies, booking calendar, blog setup.
Data scientist portfolio: projects, skills, research highlights, clean minimal design.

Reproducible training + evaluation.
Model/data versioning.
Deployment plan (batch or online).
Monitoring and regression checks.