I helped take an AI-powered records-adjudication platform from early prototype to
production with live enterprise customers. I own the service that turns messy
multi-vendor court XML into the structured data behind every automated compliance
decision.
role
Data Engineer
stack
Python · FastAPI · BigQuery
shipped
100+ merged PRs
since
Jun 2025
[01] Experience
Data Engineer
January 2026 - Present
SNH AI · Austin, TX
Own the Python/FastAPI service that transforms raw court-record XML from multiple vendors into the structured features driving automated compliance decisions.
Helped take the product from prototype to production launch, supporting multi-week client UAT and shipping 100+ merged pull requests across parsing, enrichment, validation, and decision logic.
Implemented jurisdiction-specific decision logic (offense age, dispositions, court exclusions) with pytest coverage, expanding the share of records adjudicated without manual review.
Diagnosed production incidents by tracing records through a BigQuery bronze/silver/gold pipeline, fixing parser failures and ID-mapping bugs.
Built internal React dashboard features (record search, labeling, live rule evaluation) used daily by engineers and auditors.
python fastapi bigquery gcp pytest react
Data Engineer Intern
June 2025 - January 2026
SNH AI · Austin, TX
Built data labeling workflows in Label Studio for annotating charges, dispositions, and record sources used to train and evaluate the platform's AI models.
Developed Python pipelines standardizing heterogeneous vendor XML into a unified golden dataset for model training and evaluation.
Created and validated XSD schemas to enforce consistency of incoming XML across diverse data sources.
Converted to full-time after the internship, taking ownership of the platform's core transformation service.
python xml xsd label studio
Data Analyst Intern
February 2024 - August 2024
JSoftUSA · Austin, TX
Analyzed student attendance and performance data, uncovering trends that led to a 15% increase in platform adoption.
Automated ETL workflows using SQL scripts, decreasing manual report time by 30%.
Built Tableau dashboards to track performance and resource use, improving educator efficiency.
Machine learning models (Random Forest, XGBoost, k-NN, Lasso) predicting PM2.5
concentrations across the U.S., with EDA, feature selection, and cross-validation.