Muhammad Abdullah Ghani

Muhammmad Abdullah Ghani Software Engineer linkedin.com | Github.com PROFILE Aspiring MERN Stack Developer, AI Enthusiast, and DevOps Practitioner exploring and building full-stack web applications, intelligent systems, and managing CI/CD pipelines for efficient deployment. DevOps: Proficient in containerization and orchestration with Docker and Kubernetes, version control with Git and GitHub, web server management with Apache and Nginx, and automated CI/CD pipelines using GitHub Actions. AI: AI & Full-Stack Developer specializing in NLP, data labeling, LLM evaluation, and building intelligent applications. Experienced in creating AI-driven systems, preprocessing datasets, fine-tuning models, and developing scalable web apps using MERN Bug Identification: Tested library management system by making test cases on Jira and identified bug of searching with ISBN and CardId simultaneously. Testing and Coverage: Created test cases for point of sale system and tested Statement, decision, and branch coverage of Test Cases. • • • • MERN Stack | Software Testing | Development JAVA,Python,C/C++,JavaScript | AI/ML | Devops EDUCATION National University of Computer and Emerging Sciences Aug 2022 – present Software Engineering Courses: Programming Fundamentals,Object oriented, Data Structures,Design And Analysis,Database,Requirement Engineering,Testing,Operating System • PROJECTS AI-Powered Multimodal Sentiment Analysis System Developed an AI-powered sentiment analysis system to predict emotions from images and text using a ResNet18-based CNN, achieving accurate classification across seven categories. FrontEnd: Designed a dynamic React frontend with Tailwind CSS BackEnd: Enhanced model performance by addressing class imbalance, debugging predictions, and optimizing training with PyTorch and Flask. • • QuickChat AI – Cloud-Based GenAI Customer Support Chatbot (SaaS) LLMs, RAG, Python, APIs, Data Indexing, SaaS Architecture Integrated Retrieval-Augmented Generation (RAG) to provide accurate, business-specific responses from uploaded knowledge bases (text files, documents, APIs). Built pipelines for context-aware response generation, dynamic knowledge retrieval, and scalable chatbot deployment. Fine-tuned an LLM on domain-specific customer support data to improve accuracy, tone consistency, and response reliability for small business use cases. • • • NLP Review Classification & Rating Extraction Pipeline Python, Pandas, Regex, NLP Preprocessing Automated sentiment tagging, rating extraction, text cleaning, and category mapping across 6 university factors. Performed manual + semi-automated labeling to ensure high dataset quality for downstream predictive modeling. • • CERTIFICATES Supervised Machine Learning: Regression and Classification Advanced Learning Algorithms