Senior Data Scientist at CVS Health in Los Angeles. I build the data and ML engineering underneath production models: streaming pipelines, distributed training, ML platforms, and LLM agent workflows. Previously supply-chain forecasting at Walmart.
| Data pipelines | ML platforms | LLM systems |
|---|---|---|
| Kafka → Spark → Iceberg CDC streaming with partition pruning and pre-aggregation | Distributed PyTorch training on Ray across GPUs | Pydantic AI multi-agent extraction with evidence checks and human review |
| Data contracts enforced in CI with Great Expectations | Auto-forecasting package with backtesting and retail calendars built in | Agentic RAG over BigQuery and planning reports |
| Project | What it does |
|---|---|
| copytrading Swift · Python · SQLite · OpenTelemetry |
Native macOS app that turns Discord trade calls into risk-checked Alpaca orders: LLM parsing, position limits, full tracing |
| ml-platform-demo Python · Kubernetes · CI/CD |
End-to-end ML platform: training pipelines, model serving, monitoring, and CI/CD on Kubernetes |
| safe-model-deployment Python · A/B testing |
Champion–challenger deployment with traffic splitting, sequential A/B tests, and automated rollback |
| model-observability Python · Monitoring |
Production model monitoring: PSI/KL drift detection, latency SLOs, automated alerting |
| gpu-ml-learning CUDA · Triton |
GPU programming for ML, built up through hands-on kernels in Triton and CUDA |
Los Angeles, CA · LinkedIn