Hi! I’m Bradley Emi. I’m an M.S. student in Computer Science with specialization in Artificial Intelligence at Stanford University. I do research in the Stanford Vision and Learning Lab (SVL) advised by Silvio Savarese and Amir R. Zamir. I graduated from Stanford with a B.S. in Physics in 2018.

I’m interested in reinforcement learning with a focus on real-world perception. My recent work focuses on transferring knowledge from large computer vision models to active tasks for robotics in the Gibson environment.

Teaching is also important to me. I am actively involved with the Stanford Splash organization. I also recently taught for the Summer Science Program in Boulder, CO.

Starting in Fall 2019, I’m excited to be joining the Tesla Autopilot team as a machine learning scientist!

Projects

Selected Works

Active Perception

Stanford Vision Lab

In-the-loop perception for reinforcement learning in real environments that trains fast and generalizes immediately to new unseen spaces.

PDF

Project Website

ChemType

Chemical Structure Recognition

A program that can automatically recognize and parse the structure of hand-drawn chemical molecules using machine learning and classical computer vision, even generalizing to molecules unseen in training.

PDF

GitHub

Galaxy Simulation

Jet Propulsion Laboratory

I contributed an image processing pipeline for raw Hubble Space Telescope data and multi-wavelength galaxy models to GalSim, a galaxy image simulator which includes simulation of the effects of dark energy and dark matter.

PDF

GitHub

Self-Driving Cars

Uber ATG

Machine learning algorithms for perception and motion planning for autonomous vehicles at Uber.

Contact

bemi@stanford.edu