I’m a MATS fellow, having completed cohort 9 from Jan-Mar 2026 in Berkeley, California. I am continuing my research via the MATS 9.1 extension, first out of sunny Cape Town and then out of London. If you’re in London, and interested in AI safety, let’s chat!
My current research focus is hidden propensity elicitation.
I’m generally interested in solving hard problems. I worked for CubeSpace Satellite systems for 2 years, writing embedded C code to control satellites in orbit. On of my favourite projects there was writing a NAND flash file system for storage of firmware binaries on board the satellite.
I have a MSc in Computer Science from Stellenbosch University, South Africa. For my Masters, I designed and built a robotic glove that tracks your fingers in real time, allowing you to type and control a computer without a keyboard.
I’ve previously worked at Amazon Web Services for the EC2 org in Cape Town, where I was on the team of 7 that processed logging/metrics for all of EC2. We handled more logs per minute than Twitter did tweets.
I’ve worked aboard Super Yachts in the Mediterranean for two years after high
school, competed at the Intel International Science and Engineering Fair (ISEF)
when I was 16, and interned at AWS in Cape Town while completing my
undergraduate degree in Mathematical Statistics and Computer Science. I created
and maintain the open source loadshedding API
eskom-calendar which had 60k weekly downloads at
its peak.