Deployment of a Machine Learning model with FastAPI

Python, FastAPI, REST API, PostgreSQL, SQLAlchemy, Pytest, GitHub Actions, CI/CD, Machine Learning, OpenAPI

Deploying a Machine Learning Model with FastAPI

Context

As part of a Machine Learning Engineering project, I industrialized a Machine Learning model by developing a secure and documented REST API. The objective was to make the model production-ready while applying best practices in software development, automated testing, and continuous deployment.

Objectives

• Deploy a Machine Learning model as a REST API.
• Design a professional project architecture.
• Set up a PostgreSQL database.
• Develop unit and functional tests.
• Automate tests and deployment with a CI/CD pipeline.
• Document the entire project.

Architecture

Client
   │
   ▼
FastAPI
   │
   ▼
Pydantic Validation
   │
   ▼
Machine Learning Model
   │
   ▼
PostgreSQL
(Input / Output / History)
   │
   ▼
Swagger / OpenAPI

Technologies Used

• Python
• FastAPI
• Pydantic
• PostgreSQL
• SQLAlchemy
• Pytest
• pytest-cov
• Git
• GitHub Actions
• Hugging Face Spaces
• Swagger / OpenAPI

Achievements

Model Industrialization

• Deployment of a Machine Learning model via a REST API.
• Automatic validation of incoming data with Pydantic.
• Error management and standardized responses.
• Interactive documentation automatically generated with OpenAPI.

Database

• Creation of a PostgreSQL database.
• Table modeling.
• Recording user requests.
• Prediction historization.
• Full traceability of exchanges between the API and the model.

Tests

Development of a comprehensive validation strategy:

• unit tests
• functional tests
• endpoint validation
• edge case testing
• code coverage measurement with pytest-cov

CI/CD

Automation of the development cycle:

• Git Flow
• GitHub Actions
• automatic test execution
• pre-merge validation
• environment management
• secure secret management

Documentation

• Comprehensive README
• Technical documentation
• API documentation
• Swagger / OpenAPI
• Installation and deployment instructions

Skills Demonstrated

Machine Learning Engineering

• ML Model Deployment
• Model Industrialization
• Prediction API
• Data Validation

Backend

• FastAPI
• REST API
• Pydantic
• Error Handling
• OpenAPI Documentation

Database

• PostgreSQL
• SQLAlchemy
• Relational Modeling
• Prediction Traceability

DevOps

• Git
• GitHub
• GitHub Actions
• CI/CD
• Version Control
• Secret Management

Software Quality

• Pytest
• Unit Tests
• Functional Tests
• Code Coverage
• Software Architecture

Outcome

This project allowed me to transform a Machine Learning model into a production-ready application by developing a robust, tested, and documented REST API. I also implemented best practices in Machine Learning Engineering, DevOps, and continuous integration to ensure the reliability, maintainability, and traceability of the solution.

Skills Acquired

• Python
• FastAPI
• REST API
• Pydantic
• PostgreSQL
• SQLAlchemy
• Machine Learning
• Pytest
• pytest-cov
• Git
• GitHub
• GitHub Actions
• CI/CD
• Swagger
• OpenAPI
• DevOps
• Hugging Face Spaces
• API Documentation
• Software Engineering