I am an experienced software developer in a professional setting, with most of my professional experience being as a full stack developer. I am open to contract work that puts me out of my comfort zone with a notable example being work for the Independent Electoral Commission of South Africa – where I simulated and visualized the effects of the Electoral Amendment Bill.

My research focus is multiagent reinforcement learning (MARL) in a cooperative setting. I believe that MARL has the potential to solve many problems that are currently considered as intractable in the same way that single agent reinforcement learning has done. However, MARL’s main stumbling block is poor sample efficiency which is accentuated as the number of agents increases. Some methods, such as independent learning, aim to address this but the quality of their solutions are often poor. My research aims to address these issues by using centralised training to ensure high quality solutions and then using knowledge transfer to address the issue of poor sample efficiency. My research has a deep focus on theoretical and mathematical grounding, as well as tractability.

  • Multiagent Reinforcement Learning
  • Reinforcement Learning
  • Knowledge Transfer