Project
AI Image Classification App
A full-stack web application that performs real-time image classification with a TensorFlow pipeline served via Flask.
2024
Full-stack Developer
Computer Vision · MLOps · Personal Project
Computer VisionMLOpsPersonal ProjectReactTensorFlowFlaskDockerKubernetes
AI Image Classification App
A personal initiative to explore applied machine learning in production-like environments. The application enables users to upload images and receive instant predictions backed by a custom TensorFlow model.
Architecture
- Delivered a React front end with optimistic UI updates and drag-and-drop uploads for a fluid experience.
- Exposed a Flask REST API that orchestrates preprocessing, model inference, and response caching.
- Packaged services with Docker and deployed on Kubernetes, enabling horizontal scaling and rolling upgrades.
- Wired logging, health checks, and resource dashboards to observe performance in real time.
Impact
- Demonstrated how MLOps practices accelerate iteration cycles for AI-powered features.
- Served as a blueprint for future data-centric prototypes and workshops.