
1
1
Pakistan
1 year of experience
I'm Syed Marghoob Ahmad. I'm a Python programmer and AI researcher based in Islamabad. I am currently pursuing my MS in Computational Science and Engineering at NUST while working as a Research Assistant at SINES. My current work is deeply focused on artificial intelligence, specifically developing Large Language Model hallucination detection for smart agriculture, as well as building RAG pipelines. Alongside my research and software engineering projects, I also work as a home tutor.

FinAgent Pro is an intelligent multi-agent financial automation platform that revolutionizes how small and medium businesses manage their financial operations. Built with IBM watsonx Orchestrate, it deploys six specialized AI agents that collaborate seamlessly: Expense Classifier (FinBERT-powered OCR), Invoice Generator, Fraud Analyzer (ML anomaly detection), Cashflow Forecaster, Smart Financial Assistant, and a master Orchestrator. The platform processes receipts in 8 seconds (vs. 15 minutes manually), generates invoices in 6 seconds, and detects fraud in 0.3 seconds with 95.3% accuracy. Powered by Hugging Face models (FinBERT, LayoutLMv3, Zephyr-7B) for financial domain expertise, it features voice-first interaction, predictive budget alerts (14 days early), and natural language queries. The system includes real-time expense analytics with interactive charts, comprehensive invoice tracking with payment status management, and multi-modal document understanding. Built on FastAPI (Python) backend with React/TypeScript frontend, it integrates PostgreSQL, Redis, and supports Docker deployment. FinAgent Pro saves businesses 42 hours monthly, prevents $50K in annual fraud, and scales from startups to mid-market enterprises—addressing the $127B SMB financial software market with production-ready enterprise features.
23 Nov 2025