Foundational MLOps: A Containerized ML Prediction Service with FastAPI and Docker 📦

🎯 Project Overview

This project demonstrates a core MLOps workflow: taking a trained Scikit-learn Machine Learning model, wrapping it in a FastAPI web service, and packaging the entire application into a portable Docker container.

This setup ensures that the ML model can be consistently deployed and run on any environment that supports Docker, eliminating "it works on my machine" problems.

---

🛠️ Tech Stack

• Model: Scikit-learn (Logistic Regression), packaged with joblib.
• Serving API: FastAPI (Python web framework)
• Web Server: Uvicorn (ASGI server for production)
• Containerization: Docker

---

📂 Project Structure

This structure assumes the Dockerfile was moved to the project root for the final successful build.

├── app/
│   ├── main.py            # FastAPI service (loads model, defines endpoints)
│   └── requirements.txt   # Python dependencies for the container
├── model/
│   └── trained_model.pkl  # The trained Logistic Regression model artifact
├── .dockerignore          # Files to exclude from the Docker build context
├── Dockerfile             # Instructions for building the Docker image
└── README.md              # This documentation 

---

🚀 Building and Running the Service

Prerequisites

1. Docker Desktop (or Docker Engine) must be installed and running on your host machine.
2. Your terminal must be open in the root directory of the project.

Step 1: Build the Docker Image

This command reads the Dockerfile in the root and builds the image.

docker build -t iris-ml-service:v1 .

Step 2: Run the Container

This command starts a new container instance in the background (-d), mapping the internal container port 80 to your host machine's port 8888.

docker run -d -p 8888:80 --name iris-predictor-container iris-ml-service:v1

Step 3: Test the API

Confirm the service is running and predicting by sending a sample POST request.

PowerShell Command:

# 1. Define the JSON body
$body = @{ 
    sepal_length = 5.1; 
    sepal_width = 3.5; 
    petal_length = 1.4; 
    petal_width = 0.2 
} | ConvertTo-Json

# 2. Send the request
Invoke-RestMethod -Uri http://localhost:8888/predict -Method POST -Headers @{"Content-Type" = "application/json"} -Body $body

Expected Output:

prediction_class confidence input_data                                                             
---------------- ---------- ----------                                                             
               0     0.9817 @{sepal_length=5.1; sepal_width=3.5; petal_length=1.4; petal_width=0.2}

🧹 Cleanup

Stop and remove the running container to free up system resources after testing.

1. Stop the running container:

docker stop iris-predictor-container

2. Remove the stopped container instance:

docker rm iris-predictor-container

3. Remove the image (Optional - if no longer needed):

docker rmi iris-ml-service:v1