I developed a machine learning model designed to predict agricultural yields based on multiple environmental and crop-related variables. This project aimed to enhance decision-making processes in precision agriculture by providing accurate yield forecasts and data-driven insights.

To support the deployment of this solution, I implemented a complete cloud infrastructure within a Virtual Private Cloud (VPC). The application was deployed through three distinct architectures: first, on a virtual machine (VM) with manual configuration; second, on Kubernetes (GKE) for container orchestration; and third, on Cloud Run with Cloud SQL for a fully serverless solution. A comparative analysis was conducted to identify the optimal configuration in terms of cost, performance, and maintainability.

On the user interface side, I enhanced a React-based front-end to simplify yield calculations and enable dynamic curve plotting for employees. The back-end, built with FastAPI in Python, was optimized to handle data processing efficiently. Additionally, I migrated an interactive map from Python Dash to React, leveraging data stored in a MongoDB database to provide a smoother and more responsive user experience.