Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.
Already subscribed? Sign in →
Chess Stetson came to autonomous vehicles from computational neuroscience. He earned a BA in physics from Harvard, worked at MIT Lincoln Laboratory, took an MS in neuroscience at the University of Texas Health Science Center at Houston, and completed a PhD in Computation and Neural Systems at Caltech, followed by a postdoc there — along the way appearing as a neuroscience expert on National Geographic's *Brain Games* and applying AI/ML and data-science methods for Fortune 500 companies, healthcare providers and financial services firms. In 2019 he founded dRISK, headquartered in London with a Pasadena office, to build what the company calls the first true driver's test for self-driving cars. Stetson's insight was that AV safety is dominated by edge cases — individually rare, high-risk scenarios that collectively make up most of the risk — and that these could be systematically harvested and organized. dRISK uses patented knowledge-graph technology to fuse crash records, insurance data and other sources into the most diverse collection of edge cases assembled, used to train and test (and retrain) autonomous vehicles. The work has been funded in part by a major grant from the UK's Department for Transport, and the company took investment from the Foresight Williams Technology funds in 2022. Stetson continues to lead dRISK as CEO.
No articles ingested yet for dRISK. Once the hourly news pipeline is live, every article the classifier tags as mentioning this company appears here with its one-line AI summary and sentiment.
An edge-case-focused training and testing platform for autonomous vehicles and ADAS. dRISK aggregates real-world crash, incident, and scenario data into networks of interconnected 'edge cases' that can be searched, visualized, and converted into simulation and training scenarios. By concentrating on the rare, high-risk situations that conventional datasets under-represent, the platform helps developers retrain and validate driving systems against the long tail of dangerous events. The company has reported breakthrough performance gains when retraining AVs to detect high-risk edge cases using its approach.
3 patents on file, but none with both an extractable figure and an abstract on Google Patents yet.