CI/CD for ML pipelines in ML Python - ML Experiment: Train & Evaluate

Experiment - CI/CD for ML pipelines

Problem:You have an ML pipeline that trains a model and deploys it manually. This process is slow and error-prone.

Current Metrics:Training accuracy: 90%, Validation accuracy: 88%, Deployment time: 2 hours manually

Issue:Manual deployment causes delays and risks mistakes. No automation to test, build, and deploy the ML model.

Your Task

Automate the ML pipeline with CI/CD to reduce deployment time to under 10 minutes and ensure consistent model quality.

Use GitHub Actions for CI/CD automation

Keep the same model architecture and dataset

Include automated testing of model training and evaluation

Hint 1

Hint 2

Hint 3

Solution

ML Python

name: ML Pipeline CI/CD

on:
  push:
    branches: [ main ]

jobs:
  build-and-test:
    runs-on: ubuntu-latest

    steps:
    - uses: actions/checkout@v3

    - name: Set up Python
      uses: actions/setup-python@v4
      with:
        python-version: '3.12'

    - name: Install dependencies
      run: |
        python -m pip install --upgrade pip
        pip install scikit-learn

    - name: Train and evaluate model
      run: |
        python train_and_evaluate.py

    - name: Check accuracy
      run: |
        ACC=$(python get_accuracy.py)
        echo "Model accuracy: $ACC"
        if (( $(echo "$ACC < 0.85" | bc -l) )); then
          echo "Accuracy too low, failing build."
          exit 1
        fi

    - name: Deploy model
      run: |
        echo "Deploying model..."
        # Simulate deployment step
        sleep 5
        echo "Model deployed successfully."

Added GitHub Actions workflow to automate training, testing, and deployment

Included accuracy check to fail pipeline if model quality is low

Reduced manual deployment time by automating with scripts

Results Interpretation

Before: Manual deployment took 2 hours and risked errors.
After: Automated CI/CD pipeline runs training, tests accuracy, and deploys model in under 10 minutes.

Automating ML pipelines with CI/CD improves speed, consistency, and reliability of model deployment.

Bonus Experiment

Add automated data validation and model performance monitoring to the CI/CD pipeline.

💡 Hint

Use scripts to check input data quality and track model metrics over time, triggering alerts if performance drops.