HELLO, I'M
Leading AI development at NUST with expertise in LLMs, RAG systems, and GPU-accelerated computing. Currently developing enterprise-level SaaS platforms using MERN stack. Specialized in building production-ready AI solutions with modern web technologies.
I'm Muhammad Usama, an AI Developer and Full-Stack Engineer specializing in enterprise-level AI solutions,
Large Language Models, and modern web development. Currently leading AI development at NUST while building
a comprehensive HR management SaaS platform using MERN stack architecture, demonstrating expertise across
both AI technologies and full-stack development.
My expertise spans the complete development lifecycle from AI research and GPU infrastructure to full-stack
web applications with React, Node.js, Express.js, and MongoDB. I've successfully deployed multiple
production systems including SaaS platforms, RAG-based chatbots, and enterprise automation solutions across
healthcare, legal, and HR domains. Proficient with vector databases, WebSocket implementations, and modern
deployment practices.
With a background in Electrical Engineering from Namal University Mianwali and specialized AI training from NUST,
I combine deep technical knowledge with practical full-stack development experience. Currently expanding my
skill set in MERN stack development while maintaining expertise in Flask, Django, FastAPI, and low-code platforms
like Bubble. I'm passionate about building scalable, enterprise-level solutions that integrate cutting-edge AI
with robust web technologies.
Islamabad, Pakistan
Leading AI development at NUST with extensive experience in enterprise-level AI solutions. Currently developing a comprehensive HR management SaaS platform using MERN stack. Built and deployed multiple production systems including resume sorting platforms, meeting transcription services, and intelligent chatbots. Expertise in CUDA A100 GPUs, vector databases, and RAG technologies.
Remote
Delivered enterprise-level SaaS platforms and automation solutions across healthcare, legal, and coaching domains. Built complete platforms with user authentication, role-based access, payment integration, and advanced AI chatbot functionality. Expertise spans low-code platforms (Bubble), WordPress automation, RAG systems, and production-ready deployments.
NUST, Islamabad
Completed an intensive boot camp covering topics such as fundamentals of AI, deep learning, Python, machine learning, computer vision, natural language processing, and special topics in deep learning.
Final grade: 2.98/4
Final grade: 893/1100
Currently developing a comprehensive HR management SaaS platform using full MERN stack architecture. Features complete candidate lifecycle management, dynamic test creation, protective full-screen testing with activity monitoring, AI-powered result analysis, multi-organization support with directorates, real-time WebSocket support system, and dynamic role-based permissions. A career-defining enterprise-level platform revolutionizing hiring processes.
View ProjectBuilt a comprehensive SaaS platform for intelligent resume sorting using custom rubrics and job descriptions. Features automated analysis, detailed explanations for sorting decisions, and comprehensive resume insights. Complete with user authentication, database integration, and deployed on rs.seecs.edu.pk.
View ProjectDeveloped an interactive platform where users input YouTube video URLs to ask questions about content, generate summaries, attempt quizzes, and receive automated notes. Built with advanced NLP capabilities for comprehensive video content analysis and educational enhancement.
View ProjectDeveloped a smart chatbot called "RegulAI Assistant" for organizational policy access. Leveraging Retrieval-Augmented Generation (RAG), the system efficiently addresses inquiries regarding organizational policies and regulations. Developed an API for seamless ML model deployment and integration within a Flask frontend.
View ProjectFinal Year Project that harnessed the power of Federated Learning for Cellular Traffic Prediction, improving the accuracy of predictions while preserving user data privacy. Compared performance of Federated Learning algorithms (FedAvg and FedAtt) with traditional methods like ARIMA, GARCH, Deep Neural Networks, and LSTM.
View ProjectDeveloped a Question Answering system leveraging a fine-tuned BERT-based model on the SQuAD 2.0 dataset. The system is integrated with a Flask frontend, supporting PDF, TXT, and Excel file formats. It extracts relevant information from documents, provides accurate answers to user queries, and intelligently identifies unanswerable questions.
View ProjectDeveloped "NustGuide", a comprehensive React Native mobile application designed specifically for NUST students. The app helps students solve various queries, access university services, and navigate campus resources. Currently upgrading to include voice call agent functionality for fully interactive student support.
View ProjectBuilt a comprehensive SaaS platform for German legal document analysis with ~23k legal documents. Features complete user authentication, role-based access control, and admin dashboard for vector database management. Users can query legal problems and receive answers with exact paragraph references. Admins can dynamically add documents to the vector database, with full user management capabilities and source document linking.
View ProjectDeveloped a comprehensive coaching platform on Bubble integrating AI chatbot functionality with complete user management and payment processing systems. Features dynamic one-time promo code generation where admins can provide time-limited access to customers. Demonstrates expertise in low-code development while maintaining enterprise-level functionality and AI integration capabilities.
View ProjectDeveloped a comprehensive WordPress job portal automation system for HubMed.ai. Integrated intelligent web scraping from 15+ healthcare recruitment portals with one-click job posting via WP Admin. Features custom cost-free maps integration, daily automated job scraping and posting, advanced CAPTCHA bypass, and session management. Successfully maintains 20,000+ healthcare job listings with 99.8% data accuracy and full automation.
View ProjectAn AI system that uses transfer learning to detect fish hunger patterns in aquaculture settings. The model analyzes fish behavior and movement patterns to optimize feeding schedules and reduce waste. Developed as part of freelance work on Fiverr, this solution helps fish farmers improve efficiency and sustainability in their operations.
View ProjectA machine learning-based system for detecting and preventing cyberattacks. The solution uses advanced classification algorithms to identify malicious patterns in network traffic and system behavior. Features comprehensive threat detection capabilities with high accuracy and minimal false positives, helping organizations protect their digital infrastructure from various cyber threats.
View Projectmusama2019@namal.edu.pk
(+92) 3096313137