Harvard T.H. Chan School of Public Health
Clinical Outcomes Data Pipeline
Python and SQL pipelines that clean 500K+ daily patient records, standardize VA and EHR datasets, and prepare data for regression analysis and clinical outcome modeling.
Software
Programmer
Architecting scalable data platforms, analytics pipelines, and research infrastructure.
EHR Infrastructure
Billions of clinical records
Large-scale pipelines for cleaning, standardizing, and preparing health data for research analysis.
Spatial Transcriptomics Petabyte-scale research dataDistributed processing patterns for high-volume biological datasets and computational workflows.
Quant Engineering Financial data systems at CitibankMetadata ingestion, SQL workflows, and machine learning for structured financial data operations.
Projects
Systems built around messy, high-volume scientific data.
Caltech
Distributed Spatial Transcriptomics Processing
Apache Spark and AWS workflows for efficient analysis of petabyte-scale spatial transcriptomics data, paired with Python ingestion services and retry-aware batch orchestration.
Citibank
Metadata Ingestion and Column Inference
Automated metadata ingestion, SQL integration workflows, and Python machine learning to infer column names and reduce manual data lineage documentation effort.
UC Santa Cruz Genomics Institute
Electrophysiology Batch Processing
Parallel Python pipelines for large electrophysiology datasets and clinical records, reducing runtimes from roughly 6 hours to under 2 hours while supporting 3TB+ workloads.
Experience
Selected Work
Programmer II
- Built Python and SQL pipelines to clean 500K+ daily patient records for regression analysis and clinical outcome modeling.
- Developed parallel batch-processing workflows for 1,000+ automated batches in hours.
- Processed and standardized VA and EHR datasets to support reliable downstream analysis.
- Refactored 20+ data processing scripts to improve reliability and support robust statistical inference.
Senior Software Engineer
- Built and maintained a scalable Python data analysis pipeline, reducing processing time by 30-50 percent.
- Developed Apache Spark pipelines for petabyte-scale spatial transcriptomics analysis on AWS.
- Built Python and AWS Batch ingestion services with retry logic and parallel orchestration, processing 10,000+ files per week with under 1 percent failure.
Quantitative Software Engineer
- Automated metadata ingestion and transformation in Python, reducing manual handling time by roughly 40 percent.
- Streamlined integration workflows with structured SQL procedures, reducing manual intervention by roughly 30 percent.
- Applied machine learning in Python to infer and assign column names, cutting data lineage documentation effort by roughly 25 percent.
Data Engineer
- Built parallel Python pipelines for large electrophysiology datasets, reducing processing time by roughly 40 percent and supporting 3TB+ workloads.
- Implemented parallelized batch-processing that cut runtime from about 6 hours to under 2 hours for 50,000+ clinical records per run.
- Performed dataset validation with Statsmodels and SciPy before downstream analytics.
Skills
Tools I use to move data from raw to reliable.
Languages
Python, SQL
Data Systems
Apache Spark, AWS Batch, parallel processing, ingestion pipelines
Research Data
EHR, VA datasets, spatial transcriptomics, electrophysiology, clinical records
Analysis
Regression workflows, clinical outcome modeling, Statsmodels, SciPy, machine learning
Contact