Brayden Zhang

I am a rising senior in the Engineering Science program at the University of Toronto, and work as a software engineer at Tenstorrent on distributed ML and low-latency inference.

I also research end-to-end driving, reinforcement learning, and diffusion and flow-matching models. Previously I was a machine learning intern at Ode, with Clay and LGND AI, building geospatial foundation models.

My interests are in applications of AI to the physical world — geospatial machine learning and autonomous vehicles in particular.

Research 4

RL on flow matching driving foundation models · Endpoint constrained trajectory optimization, 1st place at RealADSim · MATTERIX, Nature Computational Science · Clay Foundation Model

Tap or click an entry to expand · open the full list

RL on Flow Matching Driving Foundation Models

2026

University of Toronto, NVIDIA Research, and ELLIS Institute Tübingen · in progress

Driving policies trained end to end, combining reinforcement learning with diffusion and flow-matching policy classes.

Endpoint Constrained Trajectory Optimization for Driving Foundation Models

1st place 2025

B. Zhang, M. Golchoubian, I. Gilitschenski, B. Ivanovic, K. Chitta · RealADSim Workshop, ICCV · 1st place in the closed-loop driving challenge

Constraining trajectory optimization at the endpoint to steer a driving foundation model, evaluated closed-loop in a simulator reconstructed from real-world data rather than on logged trajectories.

MATTERIX: toward a digital twin for robotics-assisted chemistry laboratory automation

2025

K. Darvish, A. Sohal, A. Mandal, H. Fakhruldeen, N. Radulov, Z. Zhou, …, B. Zhang, et al. · Nature Computational Science 6(1), 67–82

A multiscale, GPU-accelerated simulation framework for high-fidelity digital twins of chemistry labs — robotic manipulation, powder and liquid dynamics, heat transfer and reaction kinetics, driven by a modular semantics engine.

Clay Foundation Model

2025

Clay and LGND AI

Adapted open geospatial foundation models to downstream Earth-observation tasks for Carbon Mapper and Climate TRACE, and built the evaluation around them.

Projects 4

AlphaEarth Foundations · Predicting food deserts · Solar PV forecasting · IGeo

Tap or click an entry to expand · open the full list

AlphaEarth Foundations

60K views 2025

Open implementation

A from-scratch implementation of DeepMind's embedding model for Earth observation.

Predicting Food Deserts

2024

Citadel Invitational Datathon

Modelling food access across US census tracts and the factors that predict it.

Solar PV Power Forecasting using Deep Learning

5th place 2024

ClimateHack.AI @ Harvard, Open Climate Fix

Short-horizon forecasting of solar photovoltaic output from satellite imagery sequences.

International Geography Olympiad

Bronze medal 2023

Team Canada, Royal Canadian Geographical Society

Won a Bronze medal, finished 48th internationally.

Posts 3

Courses I've Taken in University · A Collection of Cool Companies (to me) · International Geography Olympiad 2023

Courses I've Taken in University

January 2026

A Collection of Cool Companies (to me)

July 2025

International Geography Olympiad 2023

August 2023

Notes

Machine Learning · Reinforcement Learning · ML Systems · Robotics · Math · Algorithms

A linked, searchable notebook — browse all of it

Machine Learning

Primitives, probabilistic modelling, computer vision, post-training

Reinforcement Learning

Value methods, policy gradients, offline RL

ML Systems

Distributed training, inference engineering, PyTorch internals, hardware

Robotics

Robot learning, perception, state estimation, mobile robotics, manipulation

Geospatial

Earth observation, foundation models, remote sensing

Mathematics

Probability and statistics, signal analysis

Algorithms

Data structures, graphs, complexity

Systems Software

Operating systems, concurrency, memory, security