A data scientist who came to forecasting through decisions

I am a data scientist and operational researcher based in Peterborough, UK. My work sits at the point where a forecast becomes a decision: which products to reorder, which delivery slots to offer and at what price, and how to plan when demand is sparse or irregular.

From 2024 to 2026 I was the KTP Associate on a Knowledge Transfer Partnership between Priory Direct and the University of Kent, where I designed and built a forecasting and procurement decision system covering approximately 1,200 SKUs through six consecutive 30-day forecast horizons. My doctoral research at the University of Essex developed forecast-informed pricing, demand-management and routing methods for attended home delivery, published in the European Journal of Operational Research and Annals of Operations Research. In 2025 I placed sixth of 180+ participants in VN2, the first public inventory-planning competition.

My education runs from software engineering (BSc) through artificial intelligence (MSc, with Distinction) to operational research (PhD). That order explains how I work: I can build the system, I understand the models, and I judge both by the quality of the decision they support. Before moving to the UK I worked as a researcher and software developer from 2013 to 2020, focusing on natural-language processing for Arabic text and metaheuristic optimisation. I also lectured in computer science for two years.

I am revising the KTP forecasting manuscript for multi-horizon evaluation and a paper from my doctoral research. My independent ForecastOps project connects forecasting, inventory policies and supplier optimisation in a Python/FastAPI and Next.js platform. Synthetic 1,000/10,000-series exercises and PostgreSQL staging checks are complete; it is not an enterprise production deployment.

I completed the Johns Hopkins University / Great Learning Certificate Program in Agentic AI on 11 September 2026. I completed the Udemy MLOps bootcamp in April 2026. Earlier AI/NLP work focused on Arabic-text corpus preparation, annotation and ontology-based knowledge extraction, including exploratory LSTM/GRU and reinforcement-learning approaches.