I'm building production-grade AI systems that solve real-world problems. With hands-on experience in LLM integration, computer vision, and full-stack development, I turn AI from research into deployed, scalable solutions on AWS.

• 🧠 Specializing in: LLM systems (Gemini, Claude), computer vision pipelines, risk-scoring engines, data-intensive backends
• 💻 Tech Stack: Python, JavaScript, React, Express, Flask, TensorFlow, PyTorch, Docker
• 🌐 Deployment: AWS EC2, Lambda, serverless architecture, production ML pipelines
• 📚 Track Record: 15+ projects shipped, 10+ freelance clients, AWS Cloud Club co-lead
• 🎓 Status: 3rd-year CS student at FAST-NUCES (Class of 2027) — ready for internship roles

Python • YOLOv8 • EfficientNet-B0 • OpenCV | Freelance Project

Built a real-time video analytics pipeline for a youth basketball client. Fine-tuned YOLOv8 with EfficientNet-B0 binary classifier for false positive reduction. Auto-exports 20-second highlight clips with parabolic arc verification and timestamped accuracy. Production deployed.

JavaScript • Express • MongoDB • React • Gemini Vision

Production fraud detection system supporting 8+ Pakistani payment apps (EasyPaisa, JazzCash, NayaPay). Built async upload pipelines with Gemini Vision OCR, perceptual hash near-duplicate detection, and weighted risk-scoring engine. Live on AWS EC2 with PM2 + Nginx.

Python • Gemini • Gradio

3-tier hierarchical LLM memory system that compresses conversation history into summaries and extracted facts. Achieves 7.7x context extension on long sessions. Rate limiting, exponential backoff fallback, and live Gradio UI for memory inspection.

Python • Streamlit • Gemini AI

AI-powered tool converting handwritten notes (PDFs/images) to clean LaTeX and PDF output. Parallel batch processing for efficiency, diagram recognition (UML/flowcharts), and full session isolation. Perfect for students and researchers.

Python • OpenCV • Azure DevOps | Freelance Project

Production preprocessing pipeline for million-page scanned records: rotation/deskew correction, artifact removal, bleed-through suppression, and contrast normalization. Integrated with Azure DevOps CI/CD for reproducible batch execution. Reduced manual preprocessing time by ~60%.

Python • Flask • scikit-learn • TensorFlow

Full-stack ML system analyzing 105K+ articles for crisis detection. Logistic Regression + Neural Networks for 5-class categorization, TF-IDF feature extraction, A* search for multi-event risk paths. Flask REST API with interactive dashboard and model explainability outputs.

Languages: Python, JavaScript, C++, Dart, SQL, C

ML & Computer Vision: TensorFlow, PyTorch, scikit-learn, OpenCV, YOLOv8, EfficientNet, Keras

Specializations: LLM Integration (Gemini, Claude), Risk Scoring, Image Forensics, Perceptual Hashing, Production ML Pipelines

Backend: Flask, Express, FastAPI, Firebase, MongoDB, SQLite, Parquet

Frontend: React, Vite, Streamlit, Flutter

Cloud & DevOps: AWS (EC2, Lambda), Docker, PM2, Nginx, CI/CD

• 🤖 Advanced LLM systems and memory architectures
• 🎬 Computer vision pipelines for production environments
• ☁️ Scalable cloud deployment patterns on AWS
• 🚀 Open-source AI tooling & community mentorship
• 💼 Seeking AI Engineering internship roles

✅ Production Experience — 10+ freelance clients, live deployments on AWS ✅ Full-Stack Capability — Backend to frontend, ML to DevOps ✅ Real Metrics — Not toy projects; 99.2% optimization, 7.7x improvements, millions of records processed ✅ Current Skills — Cloud deployment, LLMs, computer vision, risk scoring engines ✅ Proven Learner — AWS Cloud Club co-lead, multiple certifications, hands-on with cutting-edge AI

Building production AI systems that matter. Let's create something great together. 🚀