• Integrated MSK digital pathology data pipeline with Epic healthcare system to parse and store HL7 messages securely and HIPAA-compliantly.

• Designed and implemented Rest APIs (FastAPI) for cross-team transmission of data from Postgres and

MongoDB databases.

• Built out Airflow DAGs to automate data refreshes, reducing manual intervention and ensuring daily data consistency.

• Added validation on incoming and outgoing data using Pydantic.

• Modernized developer infrastructure: moved local GitLab to Enterprise GitHub, Faust to Cloud Kafka,

changed from manual code deployments to GitHub CI/CD.

• Optimized Python + PostgreSQL backend queries, improving web UI load time from 4 s to < 1 s

• Brought on as part of a new compliance team. Modernized legacy code, ported Java processes

to Python, assisted in move from proprietary legacy databases to AWS Redshift, and ensured

backend system compliance with legal security standards.

• Optimized IoT compliance checking process; went from 3-stage manual process to ensure

devices were compliant with security standards to 1-stage automated solution, reducing lead

time dramatically.

• Mentored junior engineers on the team.

• Full stack engineer (70% backend, 30% frontend)

• Made modernization improvements to core product (Prism) -- Implemented Celery as queue-runner, better enabling multi-threading and improving throughput, helped move from AWS Redshift to Snowflake to reduce costs, implemented DB safeguards to protect PII and ensure compliance, helped improve codebase by enforcing code standards and testing, etc.

• Made frontend tweaks and improvements where needed, helped implement FE designs for core product upgrade (React FE / Flask BE)

• Wrote Python infrastructure to improve fraud-handling process.

• Facilitated move of business-critical data from on-premises legacy hardware to Amazon Redshift cluster.

• Architected, designed, and deployed graph database (Amazon Neptune) to better facilitate

anti-money-laundering services.

• Assisted in ETL process migration from loose, distributed cron-based setup to centralized

Airflow system.

• Reduced technical debt by rewriting legacy Java processes in Python.

• Developed initial product (paralegal automation as a service.)

• Python backend incorporating Tesseract for OCR and Stanford’s NLP library for string tokenization.

• App written in Django, managed client appointments within G-Suite.