Building Better Bots

It’s not every day that The AlumNUS meets someone who gets to interact with Elon Musk at work. But that’s the life of Dr Ye, a robotics machine learning engineer at Tesla - a giant of the booming electric vehicle industry helmed by the maverick billionaire. He joined Tesla in mid-2023 after a decade in various tech companies including Google. 

But if you, too, are hankering for a CV peppered with stints at big companies, Dr Ye has this advice: “Follow your passion and not what’s ‘hot’ at the moment. Trends can change quickly. It was investment banking 10 years ago, AI today and it can be different tomorrow.  Money is important but don't go just for money: you may not like what you do and might end up throwing in the towel when things get tough.” He adds, “Don’t make joining a big-name company your only goal either, for the same reasons I shared earlier.”

PASSION AND PIVOTS

For Dr Ye, who was born in China and is now a Bay Area resident, his interest in computers and programming goes back to his formative years. “I remember writing a programme in high school to format code in a way that made it look very presentable and neat,” he recalls. Projects like these made him a shoo-in for his undergraduate studies in computer science at Fudan University. He went on to pursue a PhD in the same subject at NUS.

This period marked a turning point in his passion, as he was given the liberty — and presented with the right environment — to fully satisfy his curiosity. This was nurtured by his supervisor, Dr Terence Sim, now an Associate Professor at the Department of Computer Science and a Vice Dean at the NUS Office of Admissions. Reflecting on Dr Sim’s impact, Dr Ye adds, “He would guide us and give us direction, while allowing us to decide on our research interests without micromanaging.”

Liberated by this approach, Dr Ye changed his focus from coding and programming to computer vision and machine learning (ML).  He studied and worked in that field for 10 years.  Around 2016, he started to think about how to use ML to revolutionise more fundamental sectors of the society in areas such as manufacturing, agriculture, energy, medicine. He ended up diving deep into Robotics AI.

“I wanted to connect the virtual and physical worlds and I saw the potential of ML and robotics to eliminate mundane routines and jobs from our lives,” he explains. He began exploring how ML could make robots more adaptive, dexterous, accessible and collaborative to humans, first in a more industrial setting at Google X and later in a more general setting, a humanoid, at Tesla.

TIME AT TESLA

At Tesla, he taps on ML to improve the robots that will one day provide help and support to people in all possible scenarios, from factories to homes. “The goal is to create a robot that can do anything a human can do,” he explains.

This progress involves a paradigm shift in the way we think about robots. Much of today’s conversations around such humanoids centre on their brain power, but Dr Ye highlights the importance of other human traits that machines can emulate. “Think of how you open a locked door with a key. Your eyes tell you how to put in the key but then you need to feel for how much force to use. Senses like touch are also important to instill in humanoids, so that they can help humans more effectively.”

If successful, such applications could go beyond the commercial world and help those in the domestic sphere as well. These would be a godsend for other time-crunched individuals, says Dr Ye, who is a father of two. “I would love for a robot to one day be able to fold my T-shirts and do my laundry,” he quips. 

THE NEXT UPDATE

Given that he’s immersed in the world of robots, we ask Dr Ye to cast his gaze to the future, to tell us how life might look in 2034. 

He reflects, “Hopefully, we’ll see more robots deployed in supermarkets, restaurants and factories,” he says. “The future in my mind is somehow very close to what a kids cartoon, like Tayo the Little Bus, would describe: Friendly and intelligent machines, including cars, robots and others, interact with humans in a natural, safe and effective way to significantly – 100 times or more - improve the productivity of the society as well as the wellbeing of everyone. They would ideally be powered by the learning models that have made tools like ChatGPT so useful, so that they can understand our intention and be dexterous enough to perform the tasks we want them to. This will probably become a reality, although the exact timeline is difficult to predict.”

Naturally, the conversation turns to another widely-discussed theme when it comes to robots, ML and artificial intelligence: safety and ethics. “I think it’s good that these discussions are so prominent, because it’s proof that we are paying attention to these issues and they are not blind spots,” he notes. “I think risks often arise when we do not pay attention to certain matters.” He cites COVID-19 as an example of this. Those years laid bare society’s vulnerabilities to pandemics, which some had ignored altogether. “But now that we’ve lived through it, people are a lot better prepared. I think we will see a similar trajectory for robots.”

Read more: Get advice from Dr Ning Ye on how to thrive in a tech-driven world.

Text by Keenan Pereira. Photos courtesy of Ning Ye.