https://github.com/d-kleine/mlops

Udacity - Machine Learning DevOps Engineer Nanodegree
https://github.com/d-kleine/mlops

cicd deployment mlops

Last synced: 9 days ago
JSON representation

Udacity - Machine Learning DevOps Engineer Nanodegree

• Host: GitHub
• URL: https://github.com/d-kleine/mlops
• Owner: d-kleine
• License: other
• Created: 2022-12-21T11:54:42.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2023-07-16T17:29:12.000Z (over 1 year ago)
• Last Synced: 2024-12-06T17:19:54.732Z (about 1 month ago)
• Topics: cicd, deployment, mlops
• Language: Jupyter Notebook
• Homepage:
• Size: 21 MB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md
  • Codeowners: CODEOWNERS

Awesome Lists containing this project

README

        
# Udacity Machine Learning DevOps Engineer Nanodegree

Build DevOps skills required to automate the various aspects and stages of machine learning model building and monitoring.

## Projects

### Predict Customer Churn with Clean Code

Deploying machine learning models in production:

-	PyLint and AutoPEP8

-	Git and GitHub

-	Testing with pytest and logging with logging

[Project](https://github.com/d-kleine/Udacity_MLOps/tree/main/project1_clean_code)

### Build an ML Pipeline for Short-term Rental Prices in NYC

Efficiency, effectiveness, and productivity in modern, real-world ML projects:

-	Clean, organized, reproducible end-to-end ML pipeline with MLflow

-	Track experiments, code, and results with GitHub and Weights & Biases

-	Selecting and deploying the best performing model using MLflow

[Project](https://github.com/d-kleine/Udacity_MLOps/tree/main/project2_reproducible-ml-workflow)

### Deploying a Machine Learning Model on Heroku with FastAPI

Deploying a machine learning model in Production:

-	Modeling performance, checking for bias using data cross-sections (called "slices"), and writing a model map

-	Version control of data and models with Data Version Control (DVC) 

-	Continuous Integration with GitHub Actions and Continuous Delivery/Deployment

-	Fast, type-checked and autodocumented writing of a user interface (API) with FastAPI

[Project](https://github.com/d-kleine/Udacity_MLOps/tree/main/project3_deployment-ml-pipeline)

### A Dynamic Risk Assessment System

Full automation of MLOps processes:

-	Model training and deployment

-	Establish regular assessment processes: Re-training and re-deployment of models at model drift.

-	Diagnose operational issues with models, including data integrity and stability issues, timing issues, and dependency issues 

-	Setup of automated reports for APIs

[Project](https://github.com/d-kleine/Udacity_MLOps/tree/main/project4_model-scoring-monitoring)