humanising autonomy

Understanding Vulnerable Road Users in Shoreditch, London
Humanising Autonomy exists to enable technology to be built around people, not the other way around.
We developed Behaviour AI software that analyses camera data in real-time to understand and predict pedestrian, vehicle, and passenger behaviour to better inform Advanced Driver Assistance Systems (ADAS) in vehicles - leading to safer, more efficient, or simply more pleasant experiences for all road users. Acquired in 2025 by Nextbase, global leader in the dash cam market [Press Release] [My LinkedIn Post].

The company was founded in 2017 by Leslie Nooteboom, Maya Pindeus, and myself on the premise of enabling a safer, more human-focused implementation of autonomous technology, starting with our group project at the RCA in 2016. Since then, the company has developed a methodology and software that is able to understand behaviour and intent to increase the safety of global mobility systems, as well as exploring this technology in other industries where human-machine understanding can improve human experience
Customers included >15 Fortune Global 500 enterprises including multinational leaders in the automotive and mobility space, as well as global technology companies exploring the manufacturing, smart home, and mixed reality spaces.
My contribution included:
- Product Strategy & Execution: Led the end-to-end lifecycle for the company's flagship AI product, driving the 0→1 launch: from design concept, to research, prototyping and commercialisation. Owned the tech→ product strategy, translating research into a market-leading SDK with 7 patents, >10 B2B contracts, and 100,000s of vehicles on the road using the software.
- Team Leadership & Culture: Hired, nurtured, and led a high-performing, cross-functional team of 20+ (peak of 50) product managers, designers, researchers and engineers. Fostered a culture of innovation, collaboration, and psychological safety through significant change, while managing relationships with the board and investors.
- AI & Interpretability: Designing and architecting the deployment of an automotive grade AI software. This included experience with designing and building a modular AI architecture and developing methods for the interpretability of AI model behavior to ensure safety and reliability, complying with automotive standards.
Applications of the technology