I build practical AI systems that connect intelligent reasoning with real-world software. My focus is on Agentic AI, Generative AI, Retrieval-Augmented Generation, NLP, and deep-learning solutions that solve business problems and improve user experiences.
- 🎓 Education: 6th-semester CS student at COMSATS University Islamabad, Lahore Campus.
- 🏆 Leadership: President / Chairperson at ACM COMSATS Lahore Chapter (2025–2026).
- 🚀 What I Build: Agentic AI systems, Generative AI products, NLP workflows, RAG architectures, deep-learning models (CNNs/ANNs), and AI-powered SaaS applications.
- 🧠 Currently Learning: Autonomous AI agents, advanced prompting, LLM orchestration, cloud deployment, and production-grade AI systems.
- ⚡ Fun Fact: Beyond LLMs and agents, I also train classic deep-learning models — CNNs for medical imaging and ANNs for churn prediction.
🧠 AI/ML — Agentic AI, GenAI, NLP & Deep Learning
🌐 Product UI & Frontend
⚙️ Backend, Databases & Cloud
| Project | Description | Live Link |
|---|---|---|
| Prompt Tutor AI | AI SaaS that analyzes, scores, and optimizes LLM prompts across multiple dimensions to lift output quality and reliability. | Live App |
| PhishGuard AI | NLP phishing detector — TF-IDF + Logistic Regression paired with a hybrid heuristic rules engine and a 20-worker concurrent Gmail scanner. | Live App |
| Medical Diagnostic System | Probabilistic inference engine blending Random Forest with a Bayesian Network over a Neo4j knowledge graph to rank likely conditions from symptoms. | Live App |
| Health Risk Predictor | Random Forest over a 270K-record, 10-feature lifestyle dataset predicting 3-tier health risk, served via a stateless Flask app with inline reports. | Live App |
| Project | Description | Code |
|---|---|---|
| Telco Customer Churn — ANN | Keras ANN (64→32→1) on the 7,043-record IBM Telco dataset — stratified split, inverse-frequency class weighting, and a Streamlit analytics dashboard. | GitHub |
| Brain Tumor MRI — CNN | Custom 3-block CNN (32→64→128) with Batch Normalization + Global Average Pooling classifying brain MRI scans into 4 tumor categories. | GitHub |
Auto-generated daily by a GitHub Action that “eats” my contribution graph.
