ML Platform

MLevis on Vertex AI

A standardized ML delivery framework built around Vertex AI pipelines.

The problem

Hosted-Airflow-based ML workflows carried high maintenance overhead, too many handoffs, and slow deployment cycles.

What I built / led

Moved ML orchestration into Vertex AI pipelines and standardized delivery patterns across teams.

Outcome

Résumé metrics report maintenance cost reduced by more than 80%, team handoffs reduced from 8 to 2, and deployment time reduced from 180 days to 14.

Technology

Vertex AI Pipelines · Python · CI/CD · GCP