Khush Agrawal — Product Engineer
I was previously a graduate student at CMU, working with Dr. Aviral Kumar on scaling offline RL. During the summer of 2024, I worked on autonomous construction, where I implemented a new hierarchical RL method for real-life Caterpillar dozer control, demonstrating improvements in controller time, success rate, and safety. Before graduate school, I worked on reinforcement learning (RL) at CMU as a research associate for two years, focusing on offline RL and deployable model-free RL. During my undergraduate studies, I worked at Professor Shital Chiddarwar’s IvLabs, where I built a stair-climbing robot and an assistive robot. In the summer of 2020, I worked in Professor David Held’s R-PAD Lab as part of the Robotics Institute Summer Scholars program, where I focused on deformable object manipulation.
Stackforce AI infers this person is a Robotics and Machine Learning specialist with a focus on reinforcement learning applications.
Location: Pittsburgh, Pennsylvania, United States
Experience: 3 yrs 11 mos
Skills
- Machine Learning
- Reinforcement Learning
- Robotics
Career Highlights
- Implemented hierarchical RL for real-life dozer control.
- Developed stair-climbing and assistive robots.
- Achieved significant performance improvements in RL applications.
Work Experience
AIM Intelligent Machines
Intern (3 mos)
Machine Learning Department at CMU
Graduate Assistant (2 yrs 4 mos)
Carnegie Mellon University
Research Associate (1 yr 7 mos)
Robotics Institute Summer Scholar (4 mos)
Ministry of Youth Affairs
Delegate (1 mo)
IvLabs, VNIT
Undergraduate Student Researcher (2 yrs 10 mos)
Education
Master of Science - MS at Carnegie Mellon University
Bachelor of Technology - BTech at Visvesvaraya National Institute of Technology
