About me

Hi there! I am a PhD Candidate at the Institute for Information-Oriented Control (ITR), Technical University of Munich. I am being supervised by Prof. Sandra Hirche

BackGround

I completed my Master’s from Technical University of Munich. During the time, I was also a Praktikantin at the AI Research team under Prof. Dr. Patrick Van der Smagt at Volkswagen AG. I completed my master’s thesis, focusing on the development state-space models for learning dynamics for control at the lab during this period. I was an AI Resident from 2019-2020 at Facebook AI Research (FAIR) in Menlo Park, California. I was a part of the Robotics LAB at FAIR and collaborated with Dr. Franziska Meier. During my time there, I worked on a variety of projects for improving robotic manipulation and control including representation learning for control and manipulation of robotic arms from visual data, learning loss functions for inverse dynamics training and model-based inverse reinforcement learning from visual demos.

Interests

My academic interests mostly lie in inferring more about the physical world (and creating a model of the same for control/manipulation) using machine learning architectures. Such a model has several applications - it can help answer pertinent questions regarding the causes of any phenomena that occur in their scope, help learn a policy that can output controls to drive the physical environment to a goal state that satisfies desired properties, or given a chain of previous observations, help predict what would happen next.

I am particularly interested in learning models and representations of dynamical systems that allow for easier control and manipulation. I am also interested in learning better task representations that can translate a higher level task into interpretable and easily optimizable costs automatically.

Apart from the above, I enjoy singing, doodling, reading mystery novels and have recently found a fondness for writing!