At TechVision AI, I lead the design and maintenance of Python ML pipelines for classification and recommendation systems. I built an automated Airflow retraining pipeline that cut model refresh time from 2 days to 3 hours.
I also develop and deploy ML services through REST APIs and maintain ML infrastructure on AWS. Previously, I developed backend services and anomaly-detection solutions at DataForge Solutions, where SQL and schema optimization reduced average API response time by 35%.

