Machine learning engineer focused on building reliable systems around data and model outputs.
I like work where quality is visible: explicit thresholds, useful review queues, lineage, reconciliation, and numerical validation. My projects turn messy data and technical assumptions into workflows that are testable, inspectable, and safe to change.
- SKU Sleuth: an auditable batch-classification pipeline with fail-closed quality gates, schema-drift detection, lineage, and post-load reconciliation.
- Orbital Mechanics Simulation: a NumPy and RK4 astrodynamics implementation with quantitative two-body and J2 validation, reproducible plots, and machine-readable evidence.
- Weekmark Household Lab: a privacy-first decision-support dashboard with a 13-week cash runway, exception signals, deterministic synthetic data, and documented model guardrails. Live demo
- Applied machine learning and data products
- Evaluation, data quality, lineage, and human review
- Decision-support dashboards and transparent financial modeling
- Scientific computing for aerospace and transportation
All public projects use synthetic or illustrative data. They include no employer code or private datasets.