About Me

I am an Autonomous Driving Software Engineer at NVIDIA, passionate about building safe and reliable AI systems for real-world robotics. My work focuses on simulation-based validation, safety reasoning, AI-powered evaluation, and automated triaging to improve the safety and deployment readiness of autonomous driving systems at scale.

I completed my Ph.D. in Electrical Engineering and Robotics at the University of Southern California, where I developed safety algorithms for perception-enabled robotic systems by combining control theory, motion planning, and machine learning. I was advised by Prof. Somil Bansal as part of the Safe and Intelligent Autonomy Lab and was also a visiting researcher in Aeronautics and Astronautics at Stanford University.

I enjoy building scalable tools that bridge research and production, enabling safer and more robust autonomous systems.

Research Interests: Safety of Autonomous Systems and Learning-enabled Components, Control and Planning for Safety-Critical Robots.
Research Statement

E-mail: kaustav.c.1996 [at] gmail.com

News & Updates

[June 2026] Our paper FORCE-OPT on safety evaluation of motion plans was accepted to IEEE Robotics and Automation Letters and presented at ICRA 2026 in Vienna! Read the blog here!
[March 2026] Featured in a USC alumni profile about my journey from robotics research to engineering safer, more trustworthy AI-driven systems at NVIDIA.
[February 2026] I joined NVIDIA as a full-time System Software Engineer! I am building AI-powered simulation, safety-reasoning, and automated-triaging frameworks to improve the safety and reliability of autonomous driving systems at scale.
[December 2025] I completed my Ph.D. in Electrical Engineering and Robotics at the University of Southern California! My dissertation focused on enhancing the safety of robotic systems with perception-enabled components.
[September 2025] I am honoured to be selected as a finalist for the prestigious MHI Scholars Program. Excited to present my research in the public talk competition!
[August 2025] New paper on safety evaluation of motion planning algorithms for autonomous vehicles is out on arXiv!
[May 2025] Presented our SPARQ work at ICRA 2025 in Atlanta! It was an amazing experience sharing our research on safety-aware perception failure detection.
[January 2025] SPARQ paper has been accepted at ICRA 2025! Looking forward to presenting our work on safety-aware perception failure detection in Atlanta.
[December 2024] Successfully passed my PhD qualifying exam! I am now officially a PhD candidate.
[October 2024] New paper on Enhancing Safety and Robustness of Vision-Based Controllers is out on arXiv.
[September 2024] Our paper on safeguarding AVs from perception failures is on arXiv.
[August 2024] Our paper has been accepted to IEEE Transactions on Robotics!
[May 2024] Had an amazing time at ICRA 2024 in Yokohama, Japan!
[April 2024] Looking forward to attend the 43rd Southern California Control Workshop at UC Los Angeles on April 19th!
[December 2023] Our paper on synthesizing various safety and liveness filters using Hamilton-Jacobi reachability analysis is out on arXiv!
[September 2023] New paper on detecting and mitigating system-level anomalies of vision-based controllers is out on arXiv!
[March 2022] Our paper on finding failure cases for vision-based controller using reachability analysis has been accepted to RA-L and will be presented at IROS 2023, Detroit!
[August 2022] I will be the teaching assistant for EE 482: Linear Control System for Fall 2022 at USC.
[June 2022] Our paper on obstacle aided path planning has been accepted in RA-L and will be presented at IROS 2022. Thanks to all my collaborators!
[May 2022] I passed my PhD screening exam in the Signal and Image Processing Track!