About
MLOps Mastery training program covers techniques and practices to streamline and automate the lifecycle management of ML models. It combines some aspects of DevOps with specific considerations for ML workflows to enhance the development, scalability, and reliability of deploying and managing ML models in production. Following topics will be covered: 1. Overview of MLOps: What is MLOps and its significance in the ML lifecycle. Differences between DevOps and MLOps. Key concepts like Continuous Integration (CI), Continuous Deployment (CD), Continuous Training (CT), and Continuous Monitoring (CM). Overview of the MLOps pipeline like data ingestion, model training, deployment, monitoring, and retraining. 2. Version Control: Managing different versions of datasets, model code, and configurations to track changes and ensure reproducibility. 3. Automation: Using automation tools to build, train, test, deploy, and monitor ML models. Create pipelines. 4. Model Packaging, Deployment and Serving: Packaging ML models into a deployable format and deploying them into production environments, model serving and using API for inference. Learn to create and use Docker images and introduction to Kubernetes. 5. Monitoring and Logging: Implementing monitoring solutions to track model performance, data drift, and other relevant metrics in real-time. Logging helps in debugging and understanding model behavior. Training program will cover Git and GitHub, DVC, Docker, MLFlow and Ray extensively apart from many ML libraries and tools like dask, scikit-learn, keras, Hugging Face transformers, LangChain etc. It will cover classification, forecasting and GenAI model operations. Candidates will get access to the paid OpenAI account and to the cloud environment during the training program. Minimum requirements: 1. Working knowledge of python and basic theoretical knowledge of machine learning. 2. Laptop with minimum 8GB RAM, high speed internet connection with webcam.