• Productionized recommendation models serving 12M predictions per day.
• Built training, evaluation, registry, and deployment workflows with MLflow.
• Implemented drift and performance monitoring with automated rollback thresholds.
• Cut model release time from three weeks to two days.

• Delivered production software and automation for enterprise customers.
• Improved reliability through testing, monitoring, documentation, and cross-functional delivery.
• Mentored engineers and contributed to architecture and operational reviews.

Coursework in software systems and distributed computing.

https://github.com/example/project

Python, PyTorch, MLflow, Kubernetes

Service for repeatable offline, shadow, and production evaluation of machine learning models.

https://example.com/certification