Embedded Files
I am a robotics engineer based in San Jose, California, working at the intersection of precision automation, real-time AI, and medical science automation.
My expertise spans the full stack , from designing sub-0.001mm precision assembly systems using industrial Stewart platforms, to deploying optimized deep learning models on edge hardware like the Nvidia Jetson Orin AGX with 10x inference speedups via TensorRT and CUDA kernels. I have also contributed to FDA-cleared medical device software and drafted 510(k) premarket submissions, giving me a rare combination of hands-on engineering depth and regulatory fluency under ISO 13485 and IEC 62304 standards.
I am deeply passionate about the emerging field of Physical AI, the convergence of foundation models, embodied reasoning, and real-world manipulation. I actively experiment with state-of-the-art Vision-Language-Action models including GR00T N1.5, pi0.5, piFAST, and ACT based agentic pipelines.
Beyond task performance, I am personally fascinated by a longer-horizon question: what does it mean for a robot to have a personality? In my free time, I tinker with giving robots generalizable emotions and character through AI, exploring how affect and identity can be layered into robotic behavior to make machines feel less like tools and more like genuine collaborators.
I graduated with a Master's in Robotics from the Georgia Institute of Technology (GPA: 3.85), where I worked with Dr. Ye Zhao at the Laboratory for Intelligent Decision and Autonomous Robots (lab-idar.gatech.edu ). My research focused on whole-body control of the Cassie bipedal robot combined with a soft robotic arm, using a novel two-step reinforcement learning approach inspired by ostrich locomotion.
I completed my Bachelor's in Engineering Physics with a minor in Robotics from Delhi Technological University, New Delhi. My broader research interests include reinforcement learning for legged and mobile manipulation, bio-inspired control and mechanism design, computer vision pipelines, and deploying intelligent systems in safety-critical real-world environments.
Experience
Experience
- Leadoptik Inc ( May 2024 - Present )
Robotics & Automation Engineer (Full Time)
Developed fully automated laboratory processes for assembling medical optical devices exhibiting precision < 0.001mm, using industrial Stewart platforms (PI Hexapod), reducing the average assembly time to 90 seconds from 15 minutes, resulting in a 10x reduction in assembly time.
Optimized and trained AI-based object detection models to track and characterize healthy lung tissues in real-time imaging, bringing the inference time from 80ms to 8ms using TensorRT and CUDA kernels (10x reduction).
Designed and developed an automated wafer characterization system using precision motorized stages, calculating efficiency and collecting beam profile data within 3 seconds, resulting in a 200x reduction in process time.
- Skymul Rotoye ( May 2023 - August 2023 )
Robotics Controls Intern
Designed control strategies for autonomous mobile manipulation over a Unitree B1 quadruped and a 6 DOF Unitree Z1 arm for marking areas in construction sites. Performed plane detection and segmentation through point cloud data from RGBD sensors to draw markers on arbitrary surfaces with accuracy up to 3 mm, saving labor costs and time.
Used motion and path planning algorithms like Bi-Directional RRT for manipulation, on MoveIt!, and optimizing A* for locomotion.
Used RGBD point cloud data, ROS Nav Stack, and OptiTrack motion capture system to calculate accurate marker and base transformations for localization and planning, resulting in robust navigation in a cluttered environment by generating obstacle projections of the point cloud to the local cost map. Fused the system within a behavior tree interface with exception handling to switch between missions with accompanying real-time Rviz and Gazebo visualizations, to plan the tasks appropriately.
- Laboratory for Intelligent Decision and Autonomous Robots, Georgia Institute of Technology ( August 2022 - May 2024 )
Graduate Student Researcher, Soft Manipulation Team (Supervisor- Dr. Ye Zhao)
Established high-fidelity simulation methods of representing Continuum and Soft Manipulators in popular robotic simulation environments like PyBullet and MuJoCo.
Used Inverse kinematics learnt from supervised methods for a 6DOF Soft Arm in simulation to learn robust manipulation control through imitation learning, able to track trajectories with 2 cm accuracy in simulation.
Developed Reinforcement Learning for developing whole body control of a bipedal Cassie robot equipped with a Soft Robotic Arm using a novel two-step learning approach, resulting in pelvis compensations and end effector pose reachability with around 0.13 m accuracy.
- Fluid Mechanics Laboratory, Delhi Technological University ( January 2021-June 2021 )
Student Co-Lead, Machine Learning Team (Supervisor- Prof, Raj Kumar Singh)
- DTU Altair, Delhi Technological University (September 2018-May 2022 )
Supervisor- Prof. Ranganath M. Singari
Senior Advisor - August 2021
Team Lead - June 2020- August 2021
Head of Mechanical Subsystem Design - July 2019- June 2020
Mechanical Engineer - September 2018- July 2019
Education
Education
- Georgia Institute of Technology, Atlanta, Georgia
MS, Robotics
2022-2024 - Delhi Technological University, New Delhi, India
B.Tech, Engineering Physics (minors in Robotics and Intelligent Systems)
2018-2022
Awards and Honours
Awards and Honours
Simulink Student Challenge 2021, 1st Global position, organized by Mathworks
ASME Young Engineers Paper (YEP) 2021 Finalist for paper on "Generation and Reconstruction of Turbulent Flows through Neural Network based Generative Models"
2nd International and 1st National Rank in CANSAT 2021 Competition sponsored by NASA and organized by American Astronautical Society
6th International and 1st National Rank in CANSAT 2020 Competition.
Get in touch
Get in touch
Page updated
Google Sites
Report abuse