Email: ray.mercurius@mail.utoronto.ca
Github: Ceudan
Google Scholar: Profile
Theme by orderedlist
I am a machine learning intern at Zoox (an Amazon subsidiary developing self-driving robotaxis). Previously I was a Master of Engineering in Computer Science student at Cornell University, and a machine learning research intern at Huawei Technologies Canada.
Through work experience at Autonomous Driving companies and industrial research labs I’ve gained exposure to SOTA technologies in Autonomous Robotics, Intelligent Transportation Systems, and Reinforcement Learning. Currently at Zoox I am coding up a new project that trains path planning algorithms with imitation and reinforcement learning. However instead of the typical supervised learning setup to directly imitate individual actions, we use inverse reinforcement learning to learn the reward function that humans maximize, which is much more generalized, interpretable, and modifiable. At Huawei Noah’s Ark Lab Autonomous Driving Division I investigated methods for human behavioral modelling for realistic AV simulations. Later I researched trajectory prediction algorithms, publishing papers covering topics such as diffusion models, mixture of experts, and causal feature selection. At the Data Driven Decision Making Lab and University of Toronto Transportation Research, I worked with cutting-edge techniques in traffic signal control and reinforcement learning. I have a strong mathematical foundation, having taken courses with high honors in topics such as probabilistic reasoning and statistics.
I’ve also enjoyed great extra-curricular experiences. In 2021 I rigorously trained in competitive programming. Later that year I qualified to join the University of Toronto’s International Collegiate Programming Contest Team, with 2 teammates and myself finishing 15th/83 at the 2022 East Central North America ICPC Regionals. I’ve also pursued a long-standing endeavor in quantitative finance, actively reading and investing since I was a teenager.
I have a broad set of interests and am open to new areas. These include the modelling, prediction, and planning of spatial-temporal systems which are used in robotics and transportation. I am also keen on quantitative finance, and the general application of mathematical models and machine learning.