In late 2022, Sanjil Jain, an electronic engineering graduate, was working for a B2B SaaS company in his homeland of India. His best friend, Nikhil Kumar, who had studied the same course at the Indian Institute of Technology, had bagged a job at Qualcomm. They both shared a passion for robotics – during college they were members of the robotics club and had participated in numerous hackathons – but had put their interest to one side for the sake of solid graduate careers. Then, in November that year, ChatGPT was launched, and as Sanjil puts it: “we literally lost our minds.” 

It was the moment everything changed for the pair, setting them on a path to become founders, and paving the way for them to reconnect with their dream of working in robotics. For Sanjil, ChatGPT represented the ascent of “probably the most powerful technologies of our generation” and neither were willing to sit idle while the AI revolution passed them by. Two months later, both had quit their jobs. Today, they are the co-founders of Drift, a fast-growing Bay Area start up that uses AI to enable advanced stimulation evaluations for robotics. 

Helping robotics scale at speed

Drift, which was founded in July 2025, wasn’t actually the first venture Snajil and Nikhil founded after quitting their jobs. First, they founded benny, India’s first AI search and discovery platform for ecommerce. It was a hit; the platform scaled rapidly, accumulating over 40,000 users, and was acquired two years later. “It was a pretty sweet deal,” says Sanjil. “It gave us the jumpstart we needed to start Drift.” The love for robotics they had developed at college had always lingered, he adds. “Which is why we went back to our roots; the love for building something that literally moves in the real world.”

While Drift is there to enable the rapid development of robots, it is not a hardware company itself. But as Sanjil explains, scaling simulation testing  had become a major constraint for robotics. “2025 was such an interesting year for the world in terms of technology,” he says. “Software development was solved…anyone could literally vibe code a website or an app and have fun, but robotics was still in the stone age.” In the decade since Sanjil and Nikhil had been tinkering with robots, there was little to help people speed up the pace of deployment. “We wanted to understand what is wrong with this industry and why is it not scaling at the speed of software,” says Sanjil.

Infinite scale

Sanjil and Nikhil decided to go back a layer in the development process, helping to fix the bottlenecks that result in so many robots becoming stuck in months of research and testing. Building on their understanding of the systems involved, they built an infrastructure stack that allows robotics OEMs (original equipment manufactures) to test their robots inside virtual worlds, with the ability to scale that testing to infinite levels. This is vital when a robot needs to be proven to be reliable in human environments. Sanjil gives the example of Waymo taxis. “Tens of billions of dollars have been put into scaling simulations…and they do like a hundred years of simulation in a day,” he says. “So you understand the scale that you need to make your robots reliable. Not every company has that capital or that manpower in terms of engineers to scale it, but there is a smarter way to do it now.”

And that’s where Drift comes in. “Agents can really append an existing workforce,” says Sanjil. “And when your agent has the right context, which is the most critical aspect of robotics, you can actually enable the team to deploy their robots faster, even 100X faster.” The platform builds leverage for robotics companies of every size, helping them reduce deployment time from months or years to days. “Because we know where your bottleneck is, our solution helps you test and train and evaluate your performance before you when you touch a real world use case.”

Now Drift is looking to scale. The platform is currently used by engineers from 30 countries across 200 institutions – including research labs at the likes of Stanford and MIT – helping them deploy robots 10X faster. After building it in India, and validating the concept at Antler in Singapore, they decided to shift to the US to be close to the customers it needs to really grow. It has been estimated that over the next 25 years a billion humanoid robots will be in use worldwide. Drift, says Sanjil, will enable this to happen in a decade.