Ahmed Mohammed

I’m an AI Engineer specialising in production-grade agentic AI and financial intelligence systems, with a focus on regulated-industry use cases like fraud detection, credit underwriting, and volatility forecasting.

I design and ship end-to-end ML pipelines—from raw feature engineering to containerised API deployment—built around explainability, latency, and auditability, so models can survive real fintech constraints.

In my live projects, I combine XGBoost and agentic investigation with LangGraph for real-time fraud decisioning, add a HITL checkpoint layer for regulated deployment, and deliver FastAPI scoring services packaged with Docker.

I’ve also built an explainable RAG credit risk engine on LightGBM with SHAP/LIME to generate compliance-ready rationales, and a multi-signal volatility forecasting system using EGARCH-X plus FinBERT features, with walk-forward validation and ONNX-optimised inference.