Skip to content

MLOps Project

Welcome to the MLOps project documentation! This project demonstrates modern MLOps practices for the ITMO University MLOps course.

Overview

This project showcases:

  • Data Version Control with DVC
  • ML Pipeline Development
  • CI/CD for ML Projects
  • Model Versioning and Deployment
  • Documentation as Code

Features

Data Version Control

We use DVC for managing datasets and ML models, enabling:

  • Large file versioning
  • Experiment tracking
  • Reproducible ML pipelines
  • Team collaboration

ML Pipeline

Our ML pipeline includes:

  1. Data Collection and Storage
  2. Data Processing and Feature Engineering
  3. Model Training and Validation
  4. Model Evaluation and Comparison

Development Practices

We follow industry best practices:

  • Git Flow with DVC integration
  • Automated testing and linting
  • Continuous Integration/Deployment
  • Documentation as Code