Luis Guzman

About

I enable robots to make conscious decisions based on their surroundings and to interact with the world around them. My research interests are within the scope of robot perception, specifically in the application of computer vision and reinforcement learning.

Education

2021

M.S. Robotics

University of Minnesota - Twin Cities

2020

B.S. Applied Mathematics, Engineering, and Physics

& Computer Science Certificate University of Wisconsin - Madison

Projects & Publications

2022 IEEE Robotics & Automation Magazine - Tumbling Robot Control Using Reinforcement Learning —
2022 ICRA & RA-L - Robotic Embodiment of Human-Like Motor Skills via Reinforcement Learning —
2021 Thesis - Robotic Embodiment of Human-Like Motor Skills via Sim-to-Real Reinforcement Learning
2021 - Satellite Image Building Detection using U-Net Convolutional Neural Network —
2021 - Deep Scene Relighting For Video —
2020 - Egocentric Prediction of Hand-Object Interaction —
2020 - Forecasting of COVID-19 Cases Using Time-Series Analysis of Google Mobility Data —
Past Projects

Experience

2022present

Amazon

Machine Learning Engineer

Currently, I am working in Amazon's Just Walk Out technology group. Specifically, I've been developing research tools that enable dataset generation and evaluation of new ML models on camera data from Amazon's global network of stores.

20202022

University of Minnesota

Graduate Research Assistant

During my master's program, I researched deep reinforcement learning methods to develop robotic learning-by-demonstration systems using PyTorch and MuJoCo virtual environments, which has since been published in IEEE Robotics and Automation Letters.

2021

Startup: Rover Robotics

Robotics Intern

I built the company's endurance testing process to provide data to customers about the longevity of the company's mobile robots. This required building a vision-based program that can autonomously navigate a closed course indefinitely, while monitoring the generated data to predict hardware failures. Cross-platform support for ROS1 and ROS2 was included.

2019

Dematic Corp.

Controls Engineering Intern

I designed distributed control device networks for use in automated distribution centers. During the summer intenship, I traveled to a customer site for one month to commission the systems I had helped design.

20152012

Precision Associates Inc

Automation Intern

I was an automation intern at a local manufacturing company for three summers. I worked on a variety of projects including a equipment diagnostic display, a pick-and-place robotic arm, and product inventory software.

Talks

May 2022 - ICRA - Robotic Embodiment of Human Like Motor Skills via Reinforcement Learning
November 2021 - Robots and Sensors for the Human Well-being (ROSE-HUB) IUCRC
November 2021 - Inside the Minnesota Robotics Institute: At the Heart of a Growing Robotics Industry
October 2021 - Research Seminar at HRL Laboratories

I also work on creative projects in my free time. Here are some examples of what I do for fun.