Projects

My recent work and ventures

RoboInk

AI PrintingElectronDjango
AI & Automation
AI & Automation
RoboInk: AI-Powered Personalized Printing System
Led the design and development of a personalized printing system that enables businesses to offer customized handwritten-style cards, integrating frontend, backend, automation scripts, and cloud infrastructure.

Key Features:

  • Cross-platform frontend using Electron JS for text input and print management
  • Django backend with PostgreSQL on Amazon RDS and S3 for secure storage
  • Python scripts for SVG to G-code conversion using proprietary KX software

Technologies:

Electron JS
Django
G-code
PostgreSQL
AWS
Docker

AEC Prime

AEC MarketplaceReactDjango
Full Stack
Full Stack
AEC Prime: Marketplace for AEC Professionals
Developed a marketplace tailored for professionals in the Architecture, Engineering, and Construction (AEC) industries, connecting buyers with industry-specific software tools.

Key Features:

  • Team Lead & Backend Engineer - Led project from concept to deployment
  • Django backend with Serializers and Signals for efficient data handling
  • Automated deployment pipeline with GitHub Actions and Docker

Technologies:

React
TypeScript
Django
PostgreSQL
AWS
Zustand
Chakra UI

MarkAid

AI GradingNLPEducation
AI & Education
AI & Education
MarkAid: AI-Empowered Grading Assistant
AI-empowered grading assistant for teachers, featuring auto-assessment, intelligent feedback, question-answering, and sentiment analysis.

Key Features:

  • Auto-grading for theoretical O-Level subjects (Pak-Studies exams)
  • Question-answering system to extract answers from lengthy text passages
  • Intelligent feedback system with personalized recommendations

Technologies:

Django
React
NLP
PyTorch
Apache Airflow
Docker

PointNet Classifier (PyTorch)

PointNet3D VisionPyTorch
AI & Computer Vision
AI & Computer Vision
PointNet Classifier (PyTorch)
Developed a PointNet Classifier using PyTorch for 3D object classification and semantic segmentation.

Key Features:

  • 3D object classification using ModelNet10 dataset
  • Preprocessed 3D mesh data into point clouds
  • Designed T-Net layers for alignment and transformation invariance

Technologies:

PyTorch
3D Object Classification
Semantic Segmentation
ModelNet10

Deep-Fake Audio Generation

Audio AIDeepfakeTacotron2
AI & Audio
AI & Audio
Deep-Fake Audio Generation
Research project: Generating quotes and giving them a human voice. Combined text generation with audio deepfake to create synthesized, human-like quotes.

Key Features:

  • Text generation using advanced language models
  • Audio synthesis using Tacotron2 architecture
  • Human-like voice generation for synthesized quotes

Technologies:

LSTM
Transformers
Tacotron2
Audio Processing