USC Viterbi Alumnus Works to Engineer Trust in an AI-Driven World

(Photo: Courtesy of Kaustav Chakraborty)

News

USC Viterbi Alumnus Works to Engineer Trust in an AI-Driven World

“From chatbots to autonomous cars and humanoid robots, it’s important to ensure that these technologies do not perform harmful or unintended actions for humans,” says Kaustav Chakraborty.

March 04, 2026

Aly Vander Bayse

As AI and automation continue to transform the global workforce — and our day-to-day lives — at a rapid pace, many have begun to question: How do we ensure that these systems are safe? For alumnus Kaustav Chakraborty, tackling that very issue has become the driving force of his work.

A software engineer at NVIDIA, Chakraborty focuses on enhancing the safety of robotic systems, particularly autonomous vehicles, to ensure there are adequate “guardrails” in their daily operations.

“There is always this question of trust: Do we trust whether this self-driving car will avoid crashing into someone, or that ChatGPT will not leak or misuse our information?” he told the USC Office of International Services (OIS). “I want to make sure that these automations are more trustworthy — more safe — to deal with in daily life.”

While Chakraborty first pursued electrical engineering in his undergraduate studies, he pivoted to the more hands-on field of mechanical engineering by the end of his bachelor’s program at Vellore Institute of Technology, eventually earning his master’s degree in robotics from the University of Michigan.

Robotics offered a powerful combination of intelligence and motion (systems that don’t just compute, but act), and this growing interest brought him to USC Viterbi School of Engineering’s Electrical and Computer Engineering PhD program.

“I’ve always felt drawn toward the automotive aspect of robotics, especially how self-driving cars work,” he said. “And then when the Waymos came into Los Angeles [during my studies], it was a sign that this is the future.”

At USC, Chakraborty studied under the supervision of Professor Somil Bansal, whose work focuses on enforcing safety assurances in AI-based systems — an area Chakraborty saw as both highly pressing and increasingly in demand.

“From chatbots to autonomous cars and humanoid robots, it’s important to ensure that these technologies do not perform harmful or unintended actions for humans,” he said.

The shift toward AI safety marked more than a new research direction; it became a defining academic pivot and a consequential career move. It was during this period that Bansal opened the door to a collaboration with NVIDIA Autonomous Vehicle Research Group, a highly specialized group of researchers tackling some of the most difficult challenges in self-driving technology. Together, they focused on pushing the boundaries of safety in state-of-the-art autonomous systems.

That work helped Chakraborty land an internship at NVIDIA last summer, which recently led to a full-time role. He credits USC’s mentorship, integrated opportunities and industry partnerships for enabling a “seamless transition” from academic research to real-world application.

More broadly, by having a school culture that is rooted in cross-disciplinary collaboration, rather than siloed research, students are encouraged to engage with ideas that ultimately advance their own work, according to Chakraborty.

“No one ever told me I couldn’t collaborate across areas or explore different directions. With my background in mechanical engineering and robotics, being boxed into a single field would have been difficult. That collaborative environment was essential for me,” he said.

And that ethos extends far beyond engineering. As a recipient of the Annenberg Fellowship Top Off, Chakraborty was able to connect with scholars across academic focuses, expanding his view on AI as a field shaped by communication, media and ethics.

“We met with non-Viterbi doctoral students to see if we had a common research direction … and some Annenberg students had collaborative projects on how tools from robotics can be used to enhance language-based models that are used for media purposes, such as ChatGPT,” he said.

As an international student, that openness was especially important in helping him become part of the wider Trojan Family.

“There was never a day that I was in a position where I did not find the help that I wanted. … Each and every step of the way, I felt well-prepared for any challenges,” he said. “USC has a very vibrant and close-knit international community that is brimming with resources and support.”

It’s a sentiment that didn’t end when Chakraborty graduated this past December, telling OIS that he “still feels part of USC” and actively engages in the USC alumni network through his role at NVIDIA and other pursuits.

Looking ahead, Chakraborty hopes to continue his work at the intersection of innovation and responsibility, with the long-term goal of stepping into a leadership role where he can shape safety decisions in the technologies people rely on most.

“I intend to make sure this space continues to grow and strengthen how we perceive automation in daily life,” he said.