I hold a degree in Data Science and build practical data and backend systems designed to be run, inspected, and evaluated. My work spans machine-learning evaluation, SQL analytics, Python APIs, data engineering, and production delivery.
Production application (Entra role required) | Repository | Model card | Latest release
A deployed machine-learning system for estimating seven-day NYC 311 resolution risk from creation-time information. It combines chronological evaluation, leakage controls, model-integrity checks, human-review workflows, Entra application roles, least-privilege Neon audit storage, Azure Container Apps, Key Vault, Application Insights, and an immutable container release with SBOM and provenance.
Live dashboard | API documentation | Repository | Latest release
A deployed FastAPI operations service for pipeline run tracking, data quality checks, webhook alerts, metrics, and ownership metadata. The production release runs on Azure Container Apps with Neon PostgreSQL and includes API-key-protected writes, Alembic migrations, non-root containers, CI quality gates, an SBOM, and build provenance.
Live project site | Repository
40 runnable SQL analytics challenges with validation checks, learning paths, and browser-friendly challenge pages. The exercises cover cohorts, retention, window functions, revenue analysis, and data engineering SQL patterns.
| Project | Scope | Stack |
|---|---|---|
| Data Engineering Lab | Python pipelines that ingest data, load SQLite, and write report outputs | Python, pandas, SQL, SQLite |
| SignalBoard | Full-stack productivity dashboard with authentication, database models, tests, and a live deployment | Next.js, TypeScript, Prisma, NextAuth, Vercel |
- Data science and analytics: chronological evaluation, leakage controls, calibration, drift analysis, SQL, and pandas
- Python services: FastAPI, SQLAlchemy, Alembic, pytest, and PostgreSQL
- Delivery: Docker, GitHub Actions, Azure Container Apps, release automation, and observability
- Full-stack development: TypeScript, Next.js, Prisma, and authentication