I am a computer scientist and a dad with two decades of experience in Artificial Intelligence (AI). Over the years, I’ve developed foundations across a range of functions, including core research, teaching, engineering AI systems, and steering AI R&D into product.
I am currently in search of a new role. My focus is on driving/building products harnessing AI capabilities with reliability and safety as chief concerns, and informing the AI discourse/developments in the process. Brua, a chatbot harnessing large language model and an expanding suite of safety and reliability capabilities, is an expression of this focus. Hit me up for a chat about AI at large, Brua in particular, or a coffee. I am open to opportunities.
Steering AI R&D into product
Up until end 2022, I was at Graphcore, where I was driving and contributing to the development, validation, and release of hardware agnostic general purpose software capabilities for stable training of large models in reduced precision. Most recently, I’ve had the opportunity to bring Automatic Loss Scaling from research to release/product on IPU systems, driving and working alongside a geographically distributed team (Oslo, London, Bristol, Cambridge, and wider async contributions) that works across the software stack, including research and product functions.
Stints in the industry, academia, and as an entrepreneur
Before this, I was at Telenor Research, working as a Senior Research Scientist, where I had the privilege of driving the research agenda around scaling deep reinforcement learning to real world problems. I have also been deeply involved with the Norwegian Open AI Lab to collaborate on both fundamental and applied research in deep learning and deep reinforcement learning. In addition, I teach and advise students at NTNU, UiT, and UiO, and also have close ties with the startup community in Oslo.
Going further back, I held academic research positions at the Universities of Birmingham, Manchester and Oslo. This was followed by leading the AI effort at the education technology start-up Studix, helping build technology leveraged by students to carve their own learning paths in Mathematics, powered by AI and human guidance.
Core AI research background
My core AI research background spans work in ensemble learning, deep learning, co-evolution, reinforcement learning, mechanism design, and self-organisation in decentralised multi-agent systems. My PhD (2011, University of Birmingham, UK) was about the reliable use of co-evolutionary algorithms for managing the emergent behavior of negotiating/bargaining agents.
Most of my published research can be found on Google Scholar.