Research
UT Austin and Amazon are partnering to foster a rich research community centered on key focus areas. This collaboration brings together subject-matter experts from different fields to create impactful research breakthroughs on a global scale.
Dense LEO Satellite Network Coexistence and Direct-to-Handset Communications
Research Gift – PI: Jeff Andrews
As low-Earth-orbit satellite constellations expand global connectivity, new approaches are needed to ensure efficient use of limited radio spectrum. This project develops theoretical models and practical algorithms that enable multiple satellite networks to coexist while minimizing interference. The research also investigates direct-to-cellular satellite communications, helping advance future broadband and smartphone connectivity services that can reach users in remote and underserved regions.
AI Coach: Assessing and Guiding Skilled Activity from Video
Research Gift – PI: Kristen Grauman
This project develops an AI-powered coaching system that analyzes video of people performing physical activities and provides personalized, actionable feedback. By combining information about body movement, gaze, and attention, the system seeks to assess skill level, anticipate future actions, and recommend targeted improvements. The technology has potential applications in athletic training, rehabilitation, injury prevention, and expanding access to expert coaching through mobile devices.
Learning General-Purpose Embeddings for Multi-Channel Audio
Sponsored Research Project – PI: David Harwath
Modern media platforms increasingly rely on complex audio formats such as surround sound, spatial audio, and Dolby Atmos. This project develops machine learning models that transform multi-channel audio into versatile representations that support audio understanding and quality monitoring. The resulting technology can help identify sound sources, detect audio quality issues, and improve automated analysis of rich audio environments across entertainment and media applications.
Continual Improvement of Vision-Language-Action Models with Reinforcement Learning
Research Gift – PI: Peter Stone
Vision-Language-Action (VLA) models have emerged as a powerful approach for enabling robots to understand human instructions and perform complex tasks. This project develops reinforcement learning techniques that allow VLA models to continuously improve through experience, adapting to new environments and challenges while retaining previously learned capabilities. The research aims to advance more autonomous, flexible robotic systems capable of learning throughout their operational lifetime.
To support our mutual interest in developing a sustainable, diverse talent pipeline, the UT Austin-Amazon Science Hub will offer one-year Ph.D. fellowships to UT students. These fellowships will be aligned to the academic year cycle, and recipients will be selected by the Science Hub board of advisors.
Industrial Affiliates Programs (IAPs) deliver a mutually beneficial pathway for industry and UT Austin researchers to explore fundamental research topics together.
Amazon is a proud partner of the following UT Austin Industrial Affiliate Programs:
WNCG
The mission of the Wireless Network & Communications Group is to create a collaborative environment that supports research, provides highly relevant education and opportunities, promotes technical innovation, imagination and entrepreneurship in wireless networking, communications and data sciences.
Texas Robotics
Researchers in multiple departments work to advance the capability of robotics in numerous application spaces, including social, surgical, rehabilitation, vehicles, drilling, manufacturing, space, nuclear and defense.
iMAGiNE
The iMAGiNE consortium provides tools, methodologies and knowledge for engineering the machines that support intelligent applications, from the smallest circuits to the largest systems. They work with machine learning, reasoning and understanding from cloud to edge.