Staff Software Engineer Private company
Architected and implemented MLOps automation infrastructure, including initial ML training pipelines that enabled teams to independently develop additional pipelines under technical guidance. Reduced model time-to-production from weeks to days.
Led a dataset management initiative that introduced versioning, data lineage, and data pipelines for dataset materialization, reducing time spent on data by 70%. Drove unit-wide adoption of repository organization, code quality, containerization, CI/CD, and agentic development practices, improving delivery cycles and reducing post-deployment defects.
Developed a computer vision model annotation tool that reduced annotation campaign costs by 30%, and established an annotation storage system that reduced annotation defects by 20%.
