Selected work.

AI agent workflow that automates salary slip generation, payroll validation, and employee notifications, reducing monthly payroll processing time by 90% for a mid-size HR department.

RAG over 10M+ legal documents and Llama 3.1 fine-tuning for a UK law firm with 15+ lawyers, built at SyntaxBrain. Lead AI engineer and project manager.

Multi-agent system that autonomously researches B2B prospects, qualifies leads against ICP criteria, and generates personalized outreach briefs, eliminating manual prospecting at scale.

ChatGPT-like multi-agent platform offering various stores and resources to enhance agent capabilities with advanced analytics.

Custom Model Context Protocol servers enabling modular AI workflow orchestration and reducing agent integration overhead across projects.

Educational AI platform for American High School Academy with RAG-powered chatbot for staff-student communication.

GPT-4 Vision system that identifies vehicle make, model, and year from photographs, enabling automated inventory pricing and listing generation for automotive marketplaces.

End-to-end project: fine-tuning OpenAI's Whisper-Small on Arabic Common Voice data with HuggingFace Transformers. Covers feature extraction, tokenization, training, WER evaluation, and inference deployment.

Advanced tool for Exploratory Data Analysis using natural language queries and LangChain framework with comprehensive visualizations and ML insights.

Combines AI for personalized study plans and intelligent book summarization, overcoming GPT-4's token limit constraints.