Amrood Labs
Exacto Tech - AI-Powered Quality Control Management System
Led full-stack development of an industrial IoT quality control platform with 180+ API endpoints, real-time WebSocket dashboards supporting 100+ concurrent connections, and Apache Solr processing 10,000+ inspection records daily across 15+ manufacturing facilities.
Project Overview
Exacto Tech is an industrial IoT platform for manufacturing quality control, integrating computer vision AI with real-time dashboards and edge-to-cloud data synchronization.
Led full-stack development using Next.js 12, TypeScript, Prisma, and PostgreSQL with 25+ models (inspections, defects, products, stations, shifts), implementing 180+ RESTful API endpoints and role-based dashboards (Admin, Technician, Employee) with granular permissions and audit logging.
Built Flask Python ML inference API for computer vision quality inspection with Pass/Fail classification and confidence scoring, integrating with edge cameras via dynamic IP testing and automatic camera discovery on local network.
Architected Apache Solr multi-core system (3 cores: production, staging, archive) for edge-to-cloud data sync processing 10,000+ inspection records daily with cron-based transfer, conflict resolution, and automatic error recovery with retry logic.
Created real-time WebSocket dashboard (Socket.IO, AmCharts, Recharts) supporting 100+ concurrent connections with live quality metrics, FPY (First Pass Yield) calculations, defect trend analysis, and shift performance reports serving 15+ manufacturing facilities.
Deployed multi-container Docker Compose infrastructure (Node.js, PostgreSQL, Solr, Python Flask, Caddy reverse proxy) with automated backup system (7-day daily, 4-week weekly, 6-month monthly retention) and one-command deployment.
Key Features
Flask ML Inference API (Computer Vision)
Built Flask Python API for computer vision quality inspection with Pass/Fail classification and confidence scoring, integrating with edge cameras via dynamic IP testing and automatic camera discovery on local network.
180+ RESTful API Endpoints (Full-Stack)
Led development of 180+ RESTful API endpoints using Next.js, TypeScript, Prisma, and PostgreSQL (25+ models) with role-based dashboards for Admin, Technician, and Employee roles with granular permissions and audit logging.
Apache Solr Multi-Core (10,000+ Records/Day)
Architected 3-core Solr system (production, staging, archive) for edge-to-cloud data sync processing 10,000+ inspection records daily with cron-based transfer, conflict resolution, and automatic error recovery.
Real-time WebSocket Dashboard (100+ Connections)
Created Socket.IO dashboard with AmCharts and Recharts supporting 100+ concurrent connections, live quality metrics, FPY (First Pass Yield) calculations, defect trend analysis, and shift performance reports for 15+ facilities.
Docker Compose Infrastructure with Auto-Backup
Deployed multi-container infrastructure (Node.js, PostgreSQL, Solr, Python Flask, Caddy) with automated backup system: 7-day daily, 4-week weekly, and 6-month monthly retention with one-command deployment.
Challenges & Solutions
Real-time Image Processing
Processing images in real-time with ML models while maintaining low latency was challenging.
Solution
Implemented efficient Flask API with optimized ML model integration and asynchronous processing pipelines for automated defect detection.
Microservices Orchestration
Managing multiple services including database, search engine, and APIs required robust orchestration.
Solution
Architected containerized infrastructure with Docker, implementing service orchestration and health monitoring for system reliability.
Edge Device Communication
Establishing reliable real-time communication with edge devices for live monitoring was complex.
Solution
Implemented WebSocket integration with performance metrics tracking, ensuring continuous monitoring and data visualization.
Project Gallery
Dashboard view showing document organization
Advanced search interface with filters
Dashboard view showing document organization
Advanced search interface with filters
Dashboard view showing document organization
Project Details
Timeline
Jan 2023 - Present
Company
Amrood Labs
My Role
Team Lead / Full-Stack Engineer
Tech Stack
Links
Visit Live SiteTestimonial
Muhammad Junaid delivered exceptional work on our documentation platform. His technical expertise in Next.js and database optimization significantly improved our application's performance and user experience.
Alex Thompson
CTO at Exacto Tech