Worked on Azure/azureml-examples and azureml-assets repositories to enhance deployment workflows and streamline environment configuration. Focused on flexible dependency management by removing strict version constraints in YAML environment files, which enabled smoother upgrades and reduced maintenance overhead. Refactored deployment scripts using bash and YAML to support local image inputs for binary payloads, eliminating unnecessary data downloads and improving deployment speed and maintainability. Additionally, optimized inference environments by cleaning up Conda dependencies, specifically removing azureml-defaults for Python 3.11, resulting in a reduced deployment footprint. Demonstrated a methodical approach to DevOps, scripting, and infrastructure as code across multiple Azure ML projects.