Allen Deon Saunders

This project involved creating a FastAPI with ML, utilizing libraries such as TensorFlow or scikit-learn in conjunction with Python's FastAPI framework. The repository, one of two developed, was designed to streamline the development of production ML containerized services. The available source code facilitates containerization, enabling seamless deployment on modern cloud platforms using Kubernetes or other container orchestrators.

The project involved creating a model that classifies the rank and type of wine by quality using the data generated by the source. The report shows the relationships between wine characteristics and quality for the red and white variants of the Portuguese "Vinho Verde" wine. For vineyards that produce this wine variety, a single report can change their perspective on how their winemaking process affects wine quality. Understanding the characteristics of which properties contribute to the quality of wine means being ahead of the rest. The winemaking process differs from company to company, but even the best creators always seek new insights.

This repo demonstrates how to manage multi-cloud deployments using Terraform. It uses Google Cloud Platform (GCP) and AWS as modules inside a Terraform project. In addition to Terraform, some examples of Ansible have been added.

This project involved setting up a straightforward MLOps pipeline that retrieves data from a service and merges it with a Spark database. The "cron-add.py" file located in the "data-ops" folder serves as a template for a live application. Additionally, the "spark-server" application is designed to receive authentications and execute operations once authenticated.

This project functions as a monolithic repository and is compatible with various CI/CD tools, providing a user-friendly introduction to implementing MLOps. It serves as a practical example for deploying a single ML model prediction system that undergoes training against a Spark SQL database.