Ahmad Rawn

I have a background in mathematics and statistics and am currently studying artificial intelligence at Imperial College London.

I'm interested in how intelligent systems learn structured representations, reason over complex systems, and use computation to discover new knowledge.

Contact
Research

Work in progress.

Projects
tropix
Python library for 2D tropical geometry: min-plus arithmetic, tropical polynomials, Newton polytopes, regular subdivisions, exact analytical duals, and visualisation. 49 tests.
Neural Networks from First Principles
Neural network framework built from scratch in NumPy with hand-derived backpropagation, numerical gradient checking, SGD/Momentum/Adam, regularisation experiments, and ReLU linear-region analysis.
NHS RAG Agent
RAG pipeline over NHS England A&E and RTT data. 4,254 text chunks embedded with all-MiniLM-L6-v2, indexed in FAISS, answered with Llama 3 via Groq, and exposed through Streamlit.
Bayesian A/B Testing
Bayesian conversion-rate analysis in PyMC using Beta-Binomial models and NUTS, with posterior probability of superiority, credible intervals, relative uplift, and revenue-impact uncertainty.
Ahmad Rawn
London, UK
Open to conversations about mathematics,
machine learning, and research.
amdrwnwrk@gmail.com github.com/amdrwn linkedin.com/in/ahmadrawn