Sam Kravitz

Delivered full-stack features across a clinical data platform used by treatment center teams. Built REST APIs and PostgreSQL schemas for the national haemophilia registry, and architected an AI agent that answers registry questions via RAG over patient data with grounded, citation-backed responses using LLM APIs and internal tools. Integrated LLM APIs into existing production workflows with structured tool-calling and output evaluation checks to prevent unsupported answers reaching production. Containerized services with Docker, automated build/test/deployment through CI/CD on AWS and reduced environment drift. Migrated legacy Flask data services to Django REST Framework, consolidating authentication/authorization and improving protected access to health information. Optimized PostgreSQL queries and indexes for reporting (reducing runtime by more than half) and improved reliability for nightly data loads. Extended the React frontend with reusable form components and data tables, and b

Built Python utilities and data-processing scripts to transform field-test instrument outputs into analysis-ready datasets for modeling teams. Developed internal Flask web interfaces and REST endpoints for search/annotation of field data, replacing distributed spreadsheets with a shared application. Modeled experimental data across a dozen PostgreSQL tables and wrote reporting queries to keep analyses reproducible for engineering and scientific staff. Packaged analysis environments with Docker to make the setup reproducible across workstations and onboarding faster. Hardened the data-processing pipeline with unit tests (pytest) and regression checks to catch parsing/formatting errors before outputs were used by scientists. Contributed fixes across a shared Git repository with code reviews and design discussions, aligning prototypes with team coding standards and documentation conventions.