Your First MLOps Stack
Welcome to Your First MLOps Stack course! I’m Stefano, and I’m excited to guide you into the world of MLOps. We’ll cover three key stages: Data Processing: We’ll start with Apache Beam, working locally and then scaling to Google Cloud …
Overview
Welcome to Your First MLOps Stack course! I’m Stefano, and I’m excited to guide you into the world of MLOps. We’ll cover three key stages:
Data Processing: We’ll start with Apache Beam, working locally and then scaling to Google Cloud Platform, where we’ll use Dataflow for powerful data pipelines.
Model Training: You’ll learn to run Kubeflow locally and on VertexAI, mastering the essentials of creating and managing Kubeflow pipelines in the cloud.
Model Tracking: We’ll use MLflow to track your experiments, both locally and on the cloud, while diving into cloud setup with concepts like VPC peering and firewall management.
Each stage begins with a theoretical briefing before we dive into code. By the end, you’ll grasp the MLOps mindset, know which tools to use, and be ready to work effectively with major MLOps tools.
Curriculum
Curriculum
- 4 Sections
- 51 Lessons
- Lifetime
- Data Processing4
- Apache Beam and Dataflow20
- 2.1Chapter introduction: Apache Beam and Dataflow1 Minute
- 2.2An Introduction to Apache Beam4 Minutes
- 2.3The Anatomy of a Beam Pipeline6 Minutes
- 2.4Running a Beam pipeline locally7 Minutes
- 2.5Run your first Beam pipeline locally4 Minutes
- 2.6The Combine Function (CombineFn)3 Minutes
- 2.7The CombineFn in code5 Minutes
- 2.8The CombineFn pipeline on our laptop2 Minutes
- 2.9An Introduction to Window functions3 Minutes
- 2.10A look at the WindowFn code2 Minutes
- 2.11Create your GCP account1 Minute
- 2.12Subscribe to GCP3 Minutes
- 2.13A Tour on GCP5 Minutes
- 2.14Running a Beam Pipeline on Cloud: a Look at the Code4 Minutes
- 2.15Setup of GCP Service Account8 Minutes
- 2.16Let’s Run our Beam Pipeline on Dataflow12 Minutes
- 2.17An Introduction to Dataflow Templates5 Minutes
- 2.18The Code for Creating Flex Templates12 Minutes
- 2.19Flex Template in Action15 Minutes
- 2.20Two More Pipelines to Play With and Recap10 Minutes
- Model Training13
- 3.1Introduction to MLOps Training Chapter1 Minute
- 3.2The Model Development Workflow4 Minutes
- 3.3The Model Development Workflow from the MLOps Perspective11 Minutes
- 3.4A Soft Introduction to Kubeflow and VertexAI4 Minutes
- 3.5The Anatomy of a KFP Pipeline3 Minutes
- 3.6Let’s Run our First KFP Pipeline Locally (Optional)3 Minutes
- 3.7Running a KFP Pipeline Locally (Optional)4 Minutes
- 3.8How We Can Run Custom Code in KFP7 Minutes
- 3.9Let’s Run a Custom KFP Pipeline Locally5 Minutes
- 3.10A Tour in VertexAI2 Minutes
- 3.11Introduction to the VertexAI Pipeline23 Minutes
- 3.12Let’s Run Our First VertexAI Pipeline6 Minutes
- 3.13Further KFP Components: Create a Virtual Machine From KFP9 Minutes
- Experiment Tracking14
- 4.1Introduction to Experiment Tracking1 Minute
- 4.2What is Experiment Tracking?2 Minutes
- 4.3What and Why Should We Track Experiments?6 Minutes
- 4.4A Soft Introduction to MLflow and Tracking Tools4 Minutes
- 4.5What Will We Learn from MLflow?1 Minute
- 4.6The Anatomy of an MLflow Code3 Minutes
- 4.7A Look at the Code for a Local Run with MLflow4 Minutes
- 4.8Run MLflow locally10 Minutes
- 4.9A Second Example in MLflow3 Minutes
- 4.10GPC Compute Engine6 Minutes
- 4.11Explanation of the Code for Building the MLflow Infra12 Minutes
- 4.12MLflow and VertexAI: Tracking Experiment on Cloud9 Minutes
- 4.13Delete Resources and Switch off instances4 Minutes
- 4.14Final Remarks, Pros and Cons, Critical Thinking7 Minutes

