Summary
Peter Rickwood is a Principal Data Scientist based in Sydney with 10+ years of experience building production-scale spatial AI and predictive models. At Nearmap he designed efficient custom semantic-segmentation architectures that boosted accuracy while cutting memory use by over 50%, enabling large, cost-effective deployment of high-performance imagery models. He has delivered end-to-end ML solutions from learned building outlines and direct-to-polygon outputs to a proof-of-concept for wireframe roof generation purely from vision. Prior roles include principal data science and actuarial model development at IAG and leading analytics teams in research and consultancy, reflecting a rare mix of academic rigor (PhD in urban planning) and hands-on engineering (UNSW medal-winning CS background). He mentors cross-disciplinary teams, translates complex statistical methods into deployable systems, and favors practical innovations that materially reduce operational cost. An interest not obvious from his title: his work frequently bridges urban planning concepts and computer vision to turn aerial imagery into actionable built-environment insights.
10 years of coding experience
10 years of employment as a software developer
Bachelor of Science (B.Sc.), Computer Science, 1st class honours. Awarded University Medal in Computer Science, Bachelor of Science (B.Sc.), Computer Science, 1st class honours. Awarded University Medal in Computer Science at UNSW Australia
Doctor of Philosophy (Ph.D.), City/Urban, Community and Regional Planning, PhD, Doctor of Philosophy (Ph.D.), City/Urban, Community and Regional Planning, PhD at University of Technology Sydney
English