AI Developer and Researcher with 4+ years of experience in deep learning, LLM fine-tuning, and multimodal AI systems. Proven track record in academic and industrial R&D, including work in fast-paced startups. Experienced in delivering impactful solutions in NLP, Computer Vision, and MLOps, with published research and hands-on expertise in deploying production-grade AI applications.
name: Hassan Javed
role: AI Developer & Researcher
location: Pakistan
expertise:
- Large Language Models (LLaMA, GPT, Qwen, DeepSeek)
- Computer Vision & Document AI
- Multimodal AI Systems
- Time Series Analysis & Anomaly Detection
- MLOps & Production Deployment
AI & Machine Learning
LLM Frameworks & Tools
Vector & Graph Databases
Data Science
Backend & APIs
Cloud & DevOps
| Domain | Technologies & Experience |
|---|---|
| LLMs | LLaMA 2/3/3.2 (7B-70B), Qwen, DeepSeek, GPT, Fine-tuning (LORA/QLORA), Distributed Inference (vLLM) |
| RAG Systems | LangChain, LlamaIndex, GraphRAG, Vector DBs (Qdrant, ChromaDB, Pinecone, FAISS) |
| Computer Vision | LayoutLMv2/v3, LiLT, ViT, Swin, YOLOv8/v9, Mask R-CNN, Diffusion Models |
| NLP & Document AI | BERT, GPT-2, OCR, Document Extraction, Handwriting Recognition |
| Time Series | LSTM, Forecasting, Anomaly Detection, Defense Systems |
| MLOps | Multi-GPU Inference, Model Optimization, Docker, Kubernetes, AWS (SageMaker, Lambda) |
Enhanced Plant Species Identification using Metadata Fusion and Vision Transformers (2025)
- Achieved 97.27% classification accuracy using ViT models with metadata fusion
- Published in CMC: Computers, Materials & Continua Journal (TechScience)
Building production-grade AI solutions that make a difference