work primarily with Python for building and maintaining machine learning systems, using NumPy, Pandas, and PyTorch for data processing, modeling, and experimentation. My focus is on writing clean, reproducible code that can move smoothly from research to production. For deployment and automation, I use Docker, Git, and CI/CD pipelines to package models and manage releases. I build scalable inference services and workflows that run reliably in real-world environments, whether batch-based or real-time. On the operations side, I implement experiment tracking, model versioning, and monitoring to ensure long-term system health. This includes performance metrics, logging, and data-drift detection, with an emphasis on stability, debuggability, and cost-aware operation.