Hi there! I’m a 2nd-year Master’s student at Carnegie Mellon University’s Robotics Institute, working with Prof. Katerina Fragkiadaki. My research focuses on enabling embodied agents to learn from internet-scale data.

Prior to joining CMU, I was working with Prof. K. Madhava Krishna at Robotics Research Center, IIIT-Hyderabad on Visual Servoing. I also worked as a Computer Vision Engineer at Clutterbot building perception systems for home robots.

During my undergrad, I contributed to JdeRobot through Google Summer of Code working with Prof. Jose Maria Canas Plaza. Additionally, I led the research team at SRMAUV, building computer vision stack for autonomous underwater vehicles.

Apart from research, I find joy in hiking and playing keyboard/guitar. Connect with me at pkatara@andrew.cmu.edu. Let’s chat!

🔥 News

  • February, 2024: Our work ODIN: A Single Model for 2D and 3D Perception accepted at CVPR 2024.
  • January, 2024: Our work Gen2Sim: Scaling up Robot Learning in Simulation with Generative Models accepted at ICRA 2024.
  • November, 2023: Serving as a reviewer in CVPR 2024, ICRA 2024, ICLR 2024.
  • November, 2023: Presented Gen2Sim in Towards Generalist Robots Workshop at CoRL 2023.
  • November, 2023: Volunteered at CoRL 2023.
  • August, 2022: Joined Carnegie Mellon University for MS in Robotics.
  • July, 2021: Our work RTVS accepted at IRoS 2021.
  • Aug, 2020: Our work DeepMPCVS accepted at CoRL 2020.
  • Jan, 2020: Joined Robotics Research Center, IIIT-H to pursue research on vision-based navigation.
  • October, 2019: Attended Google Summer of Code Mentor Summit in Munich, Germany.
  • May, 2018: Joined JdeRobot - Universidad Rey Juan Carlos for Google Summer of Code program.
  • May 2017: Joined student research team at SRMAUV building autonomous underwater vehicle.

📝 Publications

CVPR 2024

ODIN: A Single Model for 2D and 3D Perception

Ayush Jain, Pushkal Katara, Nikolaos Gkanatsios, Adam W. Harley, Gabriel Sarch, Kriti Aggarwal, Vishrav Chaudhary, Katerina Fragkiadaki

Website | Code | Arxiv

ICRA 2024

Gen2Sim: Scaling up Robot Learning in Simulation with Generative Models

Pushkal Katara, Zhou Xian, Katerina Fragkiadaki

Website | Code | Arxiv

  • Generation to Simulation (Gen2Sim), a method for scaling up robotic skill learning in simulation.

  • Leverages foundational generative models of image and language to scale up simulation environments, tasks and demonstrations.

📖 Educations

  • Aug 2022 - Now, Masters of Science in Robotics, Carnegie Mellon University.
  • Aug 2016 - May 2020, Bachelors of Technology in Computer Science, SRM University.