ENTERPRISE EXPERIENCE
- Established the company's single source of truth for entitlement, usage, and telemetry data by building a star-schema data mart that correlates them across ten products at a daily grain.
- Drove 11x year-over-year growth in data-product usage by integrating dozens of disparate sources into reliable ETL pipelines, serving internal stakeholders and 200+ enterprise customers.
- Built internal tools and reports for teams that had never shared data (siloed by past mergers); ran sustained user interviews resulting in internal data products with widespread adoption.
- Led DevOps (Dagster orchestration, CI/CD pipelines, Kubernetes infrastructure) and authored architectural decision records that set wide-ranging data-shape and reliability rules.
- Built full-coverage integration, end-to-end, and data-quality tests (freshness, conformity) with KPI and uptime reporting.
- Designed loads to be idempotent and re-runnable, with incremental processing and backfills.
- Optimized queries by tuning partitioning and clustering, repeatedly cutting scan volume and compute cost by more than half.
FREELANCE
- Build testing harnesses for AI models and hand-evaluate model outputs against project rubrics.
- Review and rewrite LLM-written code reviews of open-source pull requests.
- Built and consulted on tools and full-stack web presences for small businesses.
- Projects included colonist.io, a live multiplayer game; designed and programmed maps and UI features in TypeScript.
TEACHING
- Taught a biweekly game-development course (pygame) to high schoolers over Zoom; authored, administered, and graded competitive algorithm problems on HackerRank.
- Coached high schoolers to high SAT and IB scores; translated postgraduate papers for English-language journals.
- Created curricula and programming, ran game-design programs, and supervised other instructors.
PROJECTS
Sandbox Chess • multiplayer real-time chess app; Rails + React with WebSockets
The Analyst • satirical LLM therapist; FastAPI + Anthropic API with a custom RAG
- Built a full-stack RAG chat app (Next.js 14/TypeScript, FastAPI on Vercel serverless, Claude Haiku 4.5) with nine public-domain Freud & Jung books embedded into a precomputed float16 matrix (rather than an external vector DB), hardened for cost.
SKILLS
Languages: Bash, HTML/CSS, Java, JavaScript, Python, Ruby, SQL, TypeScript;
Frameworks: Dagster, FastAPI, Rails, React;
Tools: Anthropic API, Claude Code, Concourse, Docker, Git, Impala, Kubernetes, Mode, PostgreSQL, Tableau