Vijay Janapa Reddi
The University of Texas at Austin, Harvard University Press, Harvard University
Papers
28
Total Citations
486
H-Index
14
About
Vijay Janapa Reddi is a pioneering researcher at the intersection of robotics, machine learning, and computer architecture, with a particular focus on resource-constrained autonomous systems and domain-specific hardware acceleration. His work addresses one of the most pressing challenges in modern robotics: bridging the gap between the heavy computational demands of intelligent autonomous machines and the physical and energy constraints they operate under. Reddi is perhaps best known for developing Air Learning, an open-source deep reinforcement learning platform for aerial robots, which has become a foundational benchmarking tool in the field, accumulating over 70 citations across its iterations. His groundbreaking "Robomorphic Computing" framework introduced a novel design methodology for hardware accelerators parameterized directly by robot morphology, addressing performance gaps of more than an order of magnitude in motion planning tasks. His F-1 roofline model further advanced the field by providing a rigorous framework for understanding compute-performance trade-offs in autonomous aerial machines. Through his "Tiny Robot Learning" initiative, Reddi has championed the deployment of machine learning on severely constrained nano-scale robots, opening new frontiers for edge AI. With cumulative citations exceeding 300 across his top works, his research is shaping the future of efficient, intelligent robotic computing systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Accelerating Robot Dynamics Gradients on a CPU, GPU, and FPGA39 citations · 2021
- 4
- 5
- 6
- 7
- 8Tiny Robot Learning (tinyRL) for Source Seeking on a Nano Quadcopter26 citations · 2021
- 9
- 10Analyzing and Improving Fault Tolerance of Learning-Based Navigation Systems22 citations · 2021