Client Portal

Overview

Built a full-stack client portal using React, Tailwind CSS, and FastAPI to empower industrial vision system users with real-time data insights. Designed for both clients and internal teams to seamlessly manage massive datasets, monitor system health, and streamline model workflows.

Core Components

Dataset Management: Upload, tag, and organize large image collections for training and analysis.
System Management: Real-time health checks, performance metrics, and server logs dashboard.
Model Management: Deploy, version, and evaluate ML models with rich performance charts.
Image Explorer: Browse, search, and filter image gallery with overlay annotations.
Inspection Assistant: Interactive workflow to identify and review missed detections for continuous improvement.

Key Features

Dynamic filtering and search for instant data retrieval
WebSocket-powered live statistics and alert notifications
Interactive image gallery with overlay annotations and export options
Secure JWT authentication with granular role-based access control

Architecture & Tech Stack

Front-End: React, Tailwind CSS, Ant Design

Back-End: FastAPI, PostgreSQL, SQLAlchemy

Real-Time: WebSockets, async Python

Deployment: Docker Compose, CI/CD pipelines

User Interface & UX

Crafted in a dark-mode theme with a grey base and light-blue accents for reduced eye strain and intuitive navigation, ensuring a polished, professional look.

Future Plans

Custom client dashboards and advanced reporting widgets
Audit logging and extended role-based permissions
Integration with external systems (PLCs, ERP, cloud storage)

Source Code

Unavailable