I am Varun Raghavendra, a Physical AI Engineer and roboticist based in Bengaluru, India. My journey in technology began with a Bachelor’s degree in Electronics and Communication Engineering from Ramaiah University, Bengaluru. Driven by a desire to push the boundaries of autonomous systems, I pursued a Master’s in Robotics at Northeastern University in Boston, where I also worked as a Research Associate at the Institute for Intelligent Networked Systems, focusing on connected swarm robotics.

Earlier, I worked as a researcher at the AI and Robotics Laboratory at IISc Bengaluru, where I contributed to autonomous unmanned aerial systems. My experience across research and industry eventually led me to entrepreneurship. I founded Vaydyn, a deep-tech startup at ARTPARK, IISc, where we built Omnipilot, India’s first ultra-low-power autopilot for UAVs. The technology was awarded a USD 1.5 million grant in 2022.

In 2024, I continued to bridge research and industry as India’s first Entrepreneur-in-Residence at Ati Motors, a privately held Physical AI company.

Today, my work spans autonomous mobile robots, UAVs, assistive robots, GNSS-denied navigation, and digital twins. Alongside my engineering pursuits, I founded the Robotics India Community and currently lead one of India’s largest scientific AI communities at Bharat1.ai, in partnership with NVIDIA, where we are building a state-of-the-art Physical-Digital AI City.

What excites me most about Physical AI is the possibility of creating embodied agents and robots that people can genuinely live and work with—systems that understand the context of conversations naturally rather than simply waiting for rigid voice commands.

Imagine a system that remembers what matters to you, understands how your day is going, and can take over mundane, repetitive, and friction-heavy tasks without requiring constant supervision. That is where Physical AI becomes truly transformative.

Beyond personal assistance, I believe the greatest opportunities lie in safety and trust. Partial autonomous-driving technologies, for example, can reduce fatigue and distraction—two major contributors to road accidents—making everyday transportation safer and more dependable.

Another important dimension is privacy. Running autonomous intelligence directly on devices can keep sensitive personal data secure rather than continuously transmitting real-world interactions to the cloud. The combination of intelligence, autonomy, safety, and privacy makes Physical AI one of the most exciting frontiers in technology today.

My motivation came largely from what I witnessed while I was in the United States. I had a front-row view of the rapid transformation taking place across the AI landscape.

I watched foundational models accelerate at an extraordinary pace, saw Jensen Huang articulate the vision of Physical AI, witnessed autonomous humanoids operating across continuous factory shifts, and followed breakthroughs such as flying quadruped robots.

At the same time, observing India’s progress from afar made something very clear to me: the ecosystem required to support and scale frontier research and innovation in Physical AI was still in its early stages.

I realized that rather than immediately starting another venture with a closed team, the higher-leverage opportunity was to help build the ecosystem itself.

That became the motivation behind my decision to return to the ecosystem that gave me my foundations in robotics and contribute to building the next generation of Physical AI in India. Through the Robotics India Community and the AI community at Bharat1.ai, my focus is on creating open collaboration, stronger talent pipelines, and high-impact research environments.

True breakthroughs in AI cannot happen in isolation. They require a connected, resilient ecosystem where researchers, engineers, entrepreneurs, students, and institutions can move forward together.

All three experiences are inseparable, but without question, I have learned the most from building startups.

Building technology for the physical world is inherently unforgiving. Creating systems that must operate continuously, day after day and year after year, with uncompromising reliability is one of the ultimate engineering challenges.

In research environments at IISc and Northeastern, progress was often measured through controlled benchmarks, novel algorithms, and published research. But building a deep-tech startup like Vaydyn changed the equation completely.

Engineering an ultra-low-power UAV autopilot meant stepping directly into the unknown. It required relentless imagination, practical problem-solving, and a sharp intuition for what the future would require.

Real-world deployment leaves very little room for comfortable fallbacks. You are forced to build systems that cannot simply rely on theoretical assumptions when they encounter the unpredictability of the physical world.

That experience taught me perhaps the most important lesson of my career: when theory meets reality, engineering becomes much more than solving a technical problem—it becomes the ability to build something that works reliably in the real world.

Start early and get your hands dirty.

Build projects, participate in student competitions, and take part in as many hackathons as possible. These experiences are among the best ways to become familiar with the concepts and realities of robotics.

I would also encourage students to explore open-source projects before spending too much time trying to learn everything from textbooks. Pick a project, understand how it works, and then work backwards to learn the concepts behind it. In many cases, this is a faster and more effective way to learn.

Once you have built a strong project and developed a good understanding of the concepts involved, start looking for internships and research opportunities. Even if an early opportunity is unpaid for a short period, being in an environment where you have access to robots and experienced engineers can be incredibly valuable at the beginning of your career.

Finally, network actively. Connect with senior engineers, researchers, and industry professionals on LinkedIn. Ask questions, understand what companies are looking for, and learn what level of preparedness the industry expects.

In robotics, practical experience and exposure to real systems can make a tremendous difference.

One of the biggest misconceptions is that robotics is simply about building a physical machine or adding AI to an existing robot.

In reality, robotics is a deeply interdisciplinary field that brings together physics, mathematics, mechanical systems, electronics, sensing, control, software, machine learning, and increasingly, foundation models and reinforcement learning.

Another misconception is that robots and AI are primarily about replacing humans. I believe the more meaningful opportunity is to augment human capabilities and take over tasks that are dull, dirty, dangerous, or difficult for people to perform safely and consistently.

Physical AI should ultimately be about creating intelligent systems that help humans accomplish more, operate in environments that are unsafe for us, and open possibilities that were previously beyond our reach.

Contrary to what many might expect, becoming a top-tier robotics engineer begins with a strong foundation in physics and mathematics—not simply learning programming languages.

Programming is certainly essential, and being comfortable with multiple object-oriented programming languages can be extremely valuable. However, these skills need to be built on a deeper understanding of how physical systems work.

From there, engineers should develop strong foundations in core robotics principles, particularly sensing, manipulation, and navigation across different robot embodiments. This needs to be complemented by deep fluency in machine learning and deep learning.

Today, reinforcement learning is increasingly becoming a baseline capability for engineers working at the frontier of robotics.

Once these fundamentals are established, there are several exciting areas in which engineers can specialize, including multimodal and bio-inspired robot design, agentic software, tactile sensing, dexterous manipulation, and aerospace robotics.

The next generation of robotics will also require engineers capable of designing autonomous systems that can operate in highly complex environments—including space, where robots must function under varying gravitational conditions and extreme uncertainty.

I envision building a community of real believers and builders—people who challenge and push each other every week, not merely by publishing research papers, but by shipping real, working products that establish new frontiers.

At our core, we must build AI with a deep sense of purpose.

That means deploying robotics to take over dull, dirty, and dangerous tasks so humans do not have to perform them. It means using intelligent systems to help discover cures for some of the most complex diseases. And ultimately, it means creating technology that genuinely improves and uplifts society.

Our ambitions should extend beyond Earth as well. I want this community to contribute to autonomous systems capable of exploring the unknown, pushing the boundaries of deep-space exploration, and, one day, helping humanity map and understand our galaxy.

The goal is not simply to build better robots. It is to build an ecosystem capable of imagining—and then engineering—the future.