- π BS Computer Science graduate from FAST-NUCES, Islamabad β Gold Medalist, CGPA 3.2/4.0
- π§ I build production LLM-powered applications, multi-agent RAG systems, and computer vision pipelines
- βοΈ Currently deepening my expertise in AWS-native AI infra (SageMaker, EKS), Kubernetes, and MLOps
- π± Ship real, production-grade mobile apps in Flutter/Kotlin β not just tutorials
- πΌ 4x internships across AI research (govt sector), product AI engineering, and mobile development
- β‘ Fun fact: my Multi-Agent RAG classroom serves sub-300ms inference under 50 concurrent sessions
Languages
Python Dart Kotlin Java C++ C# SQL
AI / ML / GenAI
PyTorch TensorFlow scikit--learn HuggingFace LangChain LangGraph OpenAI Gemini
Mobile Development
Flutter Android Firebase SQLite
Cloud & MLOps
AWS Docker Kubernetes GitHubActions MLflow Redis
Tools
Built NLP pipelines for multi-modal intelligence data; trained text classification models in PyTorch; delivered evaluation reports to senior researchers.
Shipped LLM-based product features with LangChain; built Flutter/GetX mobile modules with Firebase, owning delivery end-to-end.
Trained & benchmarked deep learning models; built MVVM Flutter/Android apps with Firebase Auth, Firestore sync, FCM.
Gemini 2.5 MediaPipe Node.js Next.js Docker Kubernetes (EKS) AWS Redis
- Scalable microservices platform supporting 800+ concurrent users at 99.7% uptime
- Real-time attention monitoring via MediaPipe + AI-driven personalization β +38% engagement
- LLM-based tutoring, interactive flashcards, and podcast generation
GPT-4o LangGraph FAISS Docker AWS SageMaker
- Multi-agent virtual classroom deployed on SageMaker with sub-300ms inference under 50 concurrent sessions
- FAISS-based RAG pipeline evaluated on 600+ queries β BERTScore F1 0.86, ROUGE-L 0.73
Flutter BLoC Firebase SQLite
- 25+ screen production-grade LMS with quiz engine, RBAC, and real-time messaging
- Offline-first architecture with SQLite caching β ~40% faster load times
YOLOv8 U-Net PyTorch DVC GitHub Actions
- Combined detection + pixel-wise segmentation on the TACO dataset β mAP@0.5 0.82, IoU 0.74
- CI/CD with GitHub Actions + DVC versioning β 50% faster test cycles
Flutter GetX Firebase OpenStreetMap
- Cross-platform delivery app for 15+ partner restaurants with live driver tracking (<5s refresh)
π± Still building up my public commit history β most of my work lives in private/enterprise repos from internships. Check the Featured Projects above for what I've actually shipped!