Editor’s Note: This is part of a series profiling “Uncommon Thinkers”: inventors, scientists, technologists and entrepreneurs transforming industries and driving positive change. All six will be recognized this week at the GeekWire Gala on Dec. 6. Uncommon Thinkers is presented in partnership with Greater Seattle Partners. Read other profiles here.
Shwetak Patel isn’t a medical professional. He isn’t a licensed electrician.
But he is a self-described “naive innovator” — a quality that helps him develop out-of-the-box ideas that turn into successful startups.
“When you think about a problem on its head, you can come up with some really cool ideas,” said Patel, a renowned University of Washington computer science professor and serial tech entrepreneur.
Patel embodies what it means to be an Uncommon Thinker.
“He’s not afraid to pursue things with a weird or different approach,” said Vikram Iyer, a UW computer science professor who collaborates with Patel. “He finds questions that people aren’t asking.”
Those questions have sparked the formation of several entrepreneurial ventures built on Patel’s inventions.
Patel co-founded Zensi, which detected noise on electrical systems to monitor the energy usage of home appliances. The Seattle startup was acquired by Belkin in 2010.
He later co-founded SNUPI, which used existing power sources in homes to notify homeowners of potential hazards such as water leaks. Sears acquired the company’s technology in 2015.
More recently Patel has been immersed in medical technologies, co-founding a health monitoring company called Senosis Health that was acquired by Google in 2017. He currently leads the Health Technologies group at Google as a distinguished scientist, in addition to his work as an endowed professor at the UW and director of the Ubicomp Lab.
“He has a creative brilliance you don’t find often,” said Gregory Abowd, dean of the college of engineering at Northeastern University and Patel’s Ph.D advisor at Georgia Tech.
When Patel arrived in Seattle in 2008, his entrepreneurial ambitions were met with raised eyebrows within the academic community. He didn’t just want to publish research papers. He wanted to get innovations out in the world.
“I knew I had to shepherd this stuff through, or it was going to get stuck on the shelf,” Patel said.
In many ways, Patel was a trendsetter. Now about half the students he works with at the UW express interest in commercializing their research, and many of his fellow professors also spend time inside tech companies.
“That was never a thing that used to come up,” he said. “Students are more excited about the impact their stuff can have broadly, beyond just a paper.”
Patel’s colleagues praise his ability to bridge academia and industry, as well as his skills as a collaborator and leader.
“He’s very supportive of the people he works with,” said Iyer, whose latest project with Patel involves tiny battery-free autonomous robots.
Added Abowd: “He’s not just a geek. He’s a very nice individual.”
Patel, 41, was born in Selma, Ala., where his parents landed after immigrating from India in the 1970s. They were engineers, but ended up managing a hotel. That’s where Patel’s father taught him how to fix broken air conditioning units and jammed vending machines — providing an early foundation for Patel’s electrical and water-sensing inventions.
“I can fix anything because of all the random stuff I did in the motels with my dad,” he said.
His parents didn’t know much about new technology, but they bought Patel a computer. “And then I just figured stuff out,” he said.
Patel graduated high school in Birmingham and headed to Georgia Tech, where he earned bachelor and Ph.D. degrees in engineering.
Abowd, his advisor, had trouble finding projects that were challenging enough for Patel. But he did come up with a way to motivate his student: telling him something couldn’t be done.
“He just loves to defy the odds, and he has the tools to back it up,” Abowd said. “He can also inspire other people to do the same. It’s a rare gift.”
Between his current jobs at Google and the UW, Patel’s work schedule is packed. But he does find time for hobbies — tinkering with cars and watching sports (Patel is a huge Alabama football fan) — and family. His wife, Julie Kientz, is a professor and chair in the Department of Human Centered Design and Engineering at the UW.
Kientz has had a front-row seat to Patel’s nonstop motor that drives him to keep tinkering and inventing.
“His idea of vacation reading is downloading the car manual for one of his cars,” said Kientz, who met Patel at Georgia Tech.
Part of what makes Patel unique is a constant desire to learn and develop expertise — and then connect the dots.
“He’s able to pick up on all these skills and apply them across to different fields,” Kientz said.
Patel is drawn to topics that could benefit from an unconventional approach. For example, when he helped develop a mobile-based spirometer that measured lung function with a smartphone, some medical experts scoffed at the idea. Now there are several spirometer apps in the market.
“Sometimes when you approach a problem by being a non-expert, you just think about it so differently,” Patel said. He added: “I tell my students that just because an expert says it is not possible, doesn’t mean it’s not possible.”
Patel, a past winner of the prestigious MacArthur Genius grant and ACM Prize in Computing, said the tech industry is at an inflection point with the rise of AI that he believes is on par with the invention of the smartphone.
Patel is excited about how generative AI and large language models can accelerate the way researchers synthesize literature or develop algorithms.
“The pace of prototyping is much faster,” Patel said. “Engineering has gotten much faster now, too.”
Even AI hallucinations, when AI models produce false or misleading information, can be useful because it presents a new way of thinking, he said.
“It might not be perfectly framed, but it takes you down a path you wouldn’t have gone down,” Patel said, adding that it’s important for researchers to have guardrails and understand limitations of an AI program.
Patel predicts that AI will help software become more personalized and more equitable. And he thinks it will help sensors — cameras, speakers, touchscreens — become more powerful.
Patel said it’s important to identify what types of problems new AI-powered software tools can help solve. And he wants more collaboration between academia and industry, particularly with making sure universities and researchers get access to the computing capacity needed to train AI models.
As for students, Patel advises to stay flexible and not take a normal path. “Thinking differently and not going with the status quo,” he said.
And it’s perfectly fine to try many things out and not specialize too early, he said. “Don’t force yourself to be pigeonholed,” he said. “It’s OK to dabble.”
