CV

PokeNet: Learning Kinematic Models of Articulated Objects from Human Observations

2026 IEEE International Conference on Robotics and Automation (ICRA)

Articulation modeling enables robots to learn joint parameters of articulated objects for effective manipulation. PokeNet is an end-to-end framework that estimates articulation models from a single human demonstration without prior object knowledge, improving joint axis and state estimation accuracy by over 27% across diverse objects.

Learning Sequential Kinematic Models from Demonstrations for Multi-Jointed Articulated Objects

arXiv preprint arXiv:2505.06363

Continual Robot Skill and Task Learning via Dialogue

ICRA 2025 Workshop on Foundation Models and Neuro-Symbolic AI for Robotics

Learning AI-System Capabilities under Stochasticity

NeurIPS 2023 Workshop on Generalization in Planning

Ph.D.

Arizona State University, Tempe, AZ

Computer Engineering

Master of Science

Arizona State University, Tempe, AZ

Computer Engineering (Computer Systems)

Bachelor of Technology

National Institute of Technology, Bhopal

Electronics and Communication Engineering

Software Development Engineer

e2Serv Technologies

• Designed backend for “Employee Handling System” by utilizing Spring Framework, Java and Thymeleaf and used JPA methods for database connections.

Deep Learning Intern

Swaayatt Robots

• Designed and developed a Fast R-CNN network to detect and recognize traffic signs in clustered environment of India with an accuracy of 87%.
• Developed an algorithm by utilizing Style Transfer and conditional GANs to convert clear environment images into rainy and dusty environment for data augmentation.

Reviewer

Conferences

Reviewer for ICRA 2026, ICRA 2025