George LindleyAI Solutions ConsultantSelected projects in analytics, data visualisation and product strategy.
Python, OpenAI, Qdrant, neural recommender
Saudi Arabia publishes about 100 government tenders a day, in Arabic, and the education ones are easy to miss. I built a tool that scores each new tender with a neural-network recommender and estimates its value, likely competitors and compliance risk. Its scored shortlists inform my work at Pearson, from tender links for colleagues to briefings on the KSA market.
Claude Code, DoWhy, JavaScript
Everyone learns that correlation isn't causation. Almost nobody learns what to do instead. I built a game that teaches causal inference from age 12: a coding agent writes the code, and I design the learning by playtesting every level. With the engine, characters and curriculum settled, a new level now takes minutes to generate, so the real work is playtesting.
Data consultant & developer · with UNU-INWEH and UC Davis
Invasive plant regulations were scattered across 106 jurisdictions, with no shared format. I centralised, cleaned and mapped them into one public database of 2,193 taxa, and built the web app. Hosted by United Nations University and presented at the Dresden Nexus Conference 2025.
Learning consultant · Pearson Middle East
Gulf preparatory-year programmes needed English courseware made for digital classrooms and local students, not imported print. I led the market research and business case for an interactive eBook series with regional lessons and online practice, now joined by an AI lesson generator for teachers. Pearson's best-selling English course in Africa, the Middle East and Turkey, 2020–2025.