Own CV accuracy end-to-end: measure and report metrics, investigate misclassifications, curate and label datasets, prioritize retraining, and build continuous learning pipelines. Translate findings into actionable fixes and acceptance criteria, collaborate with ML/CV engineers and customer teams, and eventually implement tooling and model retraining to reduce recurring error patterns in production.

Own and improve uptime for backend services, APIs, workers, and ML pipelines; build monitoring, alerting, auto-remediation, optimize GCP infrastructure and Postgres performance, run on-call rotation and blameless postmortems.