Our Projects

Lets Discover our Projects

Safe Rail Inspection System

Al-powered rail track monitoring using trolley-mounted cameras and real-time defect detection via computer vision.

Technology Stack

YOLOv7/PyTorch object detection + ONNX-optimized inference pipeline
Industrial-grade high-speed cameras + edge-computing GPUs deployed via Docker
Serverless lot for field data + PostgreSQL for defect analytics

Applications

Crack/obstruction detection at 160 FPS
Predictive maintenance scheduling using defect trend analysis
Automated compliance reporting for rail safety audits

Funding

Implemented with 2.5 Million PKR support from Pakistan Railways for pilot deployment on Karachi-Lahore routes.

Construction Safety Al System

Real-time PPE compliance monitoring and geofencing enforcement through computer vision and GPS tracking.

Technology Stack

PyTorch-based YOLOv8 models for PPE detection with ONNX optimization for edge deployment
GPS/GIS integration for geofencing and location tracking
Containerized FastAPI microservices for scalable deployment
PostgreSQL for structured data storage and Redis for real-time alerts and site analytics

Applications

Automated PPE compliance checks for helmets and vests
Geofencing-based intrusion alerts for restricted zones
Safety audit logs to support regulatory compliance

Funding

Developed with technical support from local industry for construction safety innovation.

Al Medical Diagnostics Platform

Deep learning systems for radiological pattern detection (ground glass opacity) and neurological disorder prediction (epilepsy) through multimodal data analysis.

Technology Stack

3D CNN models developed in PyTorch/MONAI with ONNX-optimized inference
DICOM-compliant architecture using FastAPI, Docker, and PostgreSQL
Hospital PACS integration via a custom application

Applications

Early-stage lung abnormality detection in CT scans
EEG-based seizure prediction algorithms
Unified diagnostic dashboard designed for clinicians

Funding

Developed through academic collaboration with Pakistan’s HEC and partner hospitals for clinical validation.

Camouflage Pest Detection System

Al-enhanced app for identifying hidden agricultural pests through imaging and pattern recognition.

Technology Stack

Custom model with ONNX-optimized edge deployment
App integration using a containerized microservices architecture
Geospatial PostgreSQL database powered by GCP Cloud Run for processing aerial data

Applications

Real-time pest hotspot mapping in crop fields
Precision pesticide deployment guidance
Crop health monitoring through infestation trend analysis

Funding

Developed with research support from Pakistan’s Higher Education Commission (HEC) for precision agriculture initiatives.

Automated Quality Control System

Al-powered conveyor line inspection with real-time defect detection and precision product segregation.

Technology Stack

YOLOv8/PyTorch object detection with ONNX-optimized edge inference
Industrial cameras integrated with Containerized FastAPI and Redis streaming
PostgreSQL and GCP Vertex AI for managing defect archives

Applications

High-speed automated product sorting
Surface defect classification (e.g., over-burning, deformations)
Batch compliance certification generation

Funding

Developed through technical partnerships with local manufacturing leaders for factory-floor integration.

Driver Assistance System

Al-powered behavioral monitoring combining facial expression analysis, posture detection, and traffic pattern recognition for enhanced road safety.

Technology Stack

Computer vision pipeline using PyTorch/OpenCV with ONNX-optimized models
Edge-cloud architecture leveraging Docker and Redis for real-time data streaming
Vehicle telemetry integration through cameras with PostgreSQL logging

Applications

Real-time driver fatigue/drowsiness alerts with 95% accuracy
Adaptive control recommendations based on vehicle type detection
Collision prevention via lane discipline analysis

Funding

Collaborative R&D project with $1 Million USD funding from Korean Government (MSIT) for system development and road testing.

Food Shelf Life Prediction App

Al-driven visual spoilage detection and expiration forecasting through real-time object recognition.

Technology Stack

PyTorch object detection models optimized with ONNX runtime
Streamlit web interface paired with FastAPI backend, deployed via Docker
PostgreSQL database integrated in a microservices architecture for batch processing

Applications

Consumer-level freshness checks using smartphone cameras
Retail inventory management with expiration tracking
Supply chain optimization through spoilage prediction

Funding

Developed with 0.4 Million PKR seed funding from Pakistan’s Higher Education Commission (HEC) for pilot testing.

Engagement Detection for Online Classes

Al-powered real-time attention tracking and participation analysis for virtual learning environments.

Technology Stack

Deep learning models built with PyTorch and optimized via ONNX runtime
Microservices architecture using Docker, PostgreSQL, and REST APIs
Serverless deployment with auto-scaling for live video processing

Applications

Automated attendance and focus metrics for educators
Personalized student performance dashboards
Institutional analytics to optimize curriculum design

Funding

Secured 0.5 Million PKR from the Higher Education Commission (HEC) Pakistan for prototyping and validation.

Autonomous Area Mapping System

Swarm robotics platform using SLAM for real-time 3D mapping and exploration of unstructured environments.

Technology Stack

ROS-based SLAM algorithms combined with PyTorch-trained navigation models optimized via ONNX
NVIDIA Jetson edge devices integrated with LiDAR/depth cameras, deployed using Docker
Distributed PostgreSQL/Redis architecture for effective swarm coordination

Applications

Disaster zone mapping to support rescue operations
Industrial site inspection in hazardous environments
Agricultural land surveying and crop analysis

Funding

Developed through collaborative R&D with Pakistan’s Higher Education Commission (HEC) and local industry partners.

Panorama Camera Ball for Disaster Management

Al-integrated spherical imaging system delivering 360° real-time situational awareness and hazard mapping in crisis zones.

Technology Stack

Computer vision-based panoramic image stitching
Embedded Linux edge computing with containerized Socket/REST microservices
PostgreSQL/PostGIS geospatial database combined with WAN for reliable field connectivity

Applications

Rapid structural damage assessment post-disaster
Live survivor detection in debris using thermal and visual data fusion
Dynamic evacuation route planning based on hazard heatmaps

Funding

Developed with technical and financial support from Pakistan’s HEC and local industry partners for disaster response innovation.

Brain Computer Interface (BCI) System

EEG-driven neural interfaces enabling therapeutic interventions, driver fatigue monitoring, and direct drone/device control through brainwave patterns.

Technology Stack

PyTorch-based deep learning models for EEG signal processing with ONNX optimization
OpenBCI hardware integration combined with Docker for scalable deployment
MongoDB time-series database and MQTT protocol for real-time neural data streaming

Applications

Non-invasive neurorehabilitation therapies for stroke patients
Driver drowsiness detection using EEG signatures during operation
Mind-controlled drone navigation for assistive technologies

Funding

Multi-phase R&D collaboration with Korea’s MSIT.

Featured Projects

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.