Umer Mehraj

Background

About Umer Mehraj

I am specializing in MLOps, currently contributing at CalPERS, where I streamline the deployment, monitoring, and management of machine learning models in production. Leveraging expertise in automation and infrastructure management, I craft efficient CI/CD pipelines using Git, Jenkins, Docker, Kubernetes, and Terraform, ensuring reliable ML solution delivery. My proficiency includes establishing reproducible processes for ML code, data, and model configurations, orchestrating infrastructure through Infrastructure as Code (IaC), and implementing robust model versioning, monitoring, logging, and automated retraining systems. Collaborating closely with data scientists, engineers, and cross-functional stakeholders, I deliver high-quality ML solutions aligned with strategic business objectives. Passionate about continuous learning and innovation, I actively stay abreast of the latest advancements in MLOps, driving excellence and transformative results in machine learning operations. What I Do Best- Design and implement robust ETL/ELT pipelines on serverless architectures (Glue, Lambda, EMR)- Build real-time and batch data platforms using Spark, Airflow, and Lakehouse models- Lead cloud migration initiatives and medallion architecture designs on AWS S3 with Lake Formation- Develop API-driven microservices with FastAPI and integrate 3rd-party identity data via secure, scalable pipelines- Empower business teams with actionable insights via Tableau, Power BI, and custom Snowflake views.