peek User

I’m a data engineer with 8 years building batch and streaming pipelines—11 years total in data—across energy, healthcare, and analytics SaaS. I’m deep on AWS (Glue, Lambda, Step Functions, Kinesis/MSK), Airflow, and dbt, integrating multi-source telemetry and operational data.

At Copay Health AI, I unified FHIR R4 EHR, practice-management, and payment-gateway data into a single HIPAA-compliant pipeline with strict PHI/payment isolation. I built the feature pipeline and retraining loop behind ML scoring, scoring ~25K patient accounts nightly and holding the collectability model at AUC ~0.81 through payer-mix shift while lifting recovered payments ~15% over a rules-based baseline.

I also modernized performance-critical workloads by redesigning a 30-minute batch scoring job into a streaming consumer on AWS MSK and Kinesis—cutting scoring-path lag to under 2 minutes on live payment and claims events. For “agentic billing,” I grounded Navi on curated rules and policy data using pgvector for RAG.

Before that, at IDARE, I engineered the digital-twin data pipeline for energy assets—connecting, transforming, and loading multi-source sensor/SCADA and asset-registry data into one queryable model at ~2M sensor readings/day. I’m Python-first and reliability-obsessed, with the security and compliance mindset needed for HIPAA- and PCI-aware systems.