Building a Reproducible Model Workflow Udacity

Overview

In particular, it teaches the fundamentals of MLops and how to: 

a) create a clean, organized, reproducible, end-to-end machine learning pipeline from scratch using MLflow; 

b) clean and validate the data using pytest; 

c) track experiments, code, and results using GitHub and Weights & Biases; 

d) select the best-performing model for production and; 

e) deploy a model using MLflow. 

Along the way, it also touches on other technologies like Kubernetes, Kubeflow, and Great Expectations and how they relate to the content of the class. The Building a Reproducible Model Workflow program is offered by Udacity.

Course Skills

• Machine learning configuration management
• Exploratory data analysis
• Weights & biases
• Data cleaning
• Model deployment
• Hydra
• Data versioning
• Non-deterministic data testing
• Machine learning pipeline creation
• Deterministic data testing
• Pytest
• MLflow
• Data validation
• Model testing
• Machine learning experiment tracking
• Data pre-processing for ML
• Model evaluation
• Inference pipelines
• Data splitting
• Model performance metrics