Over the last few decades, it feels like the whole world has gone digital, with music and movies, radio and television broadcasts, and communications networks converting their signals to ones and zeros. But digital signals are not the best choice for every application: In fact, critical technologies, including 6G wireless networks, satellite communication, mobile phones, radar systems, and a wide array of medical, scientific and industrial instruments and sensors, still rely heavily on analog radio-frequency (RF) signals.
Zheng Liu, who joins the Department of Electrical and Computer Engineering as an assistant professor in January 2026, is designing better, more efficient analog and radio-frequency circuits to improve all of these applications. “We are using artificial intelligence to create a better design space and push the performance of these circuits,” he says. “Instead of ones and zeroes, we are dealing with analog signals and sensors that gather data in the real physical world. This world is more complex than the pure digitized world. We need to deal with how much power a circuit generates, how much noise it generates, its speed and its bandwidth. There are a ton of design parameters that trade off of each other.”
Liu studied optical engineering as an undergraduate at Peking University before earning a master’s degree at the University of California, Los Angeles, focusing on analog and digital chip design. He then took a position at Skyworks Solutions in California, a company that develops cutting-edge analog components in smart phones for wireless communication. Working there, things crystallized for Liu and he decided to focus exclusively on radio-frequency integrated circuit design. He left Skyworks to earn a PhD at Princeton University, then took another industrial research job at Texas Instruments. For the last two years, he has worked in the company’s Dallas-based Kilby Labs (named for Nobel Prize recipient and UW-Madison alum Jack St. Clair Kilby), researching radio frequency and millimeter wave circuits and systems.
The Kilby Lab, he says, is ultimately focused on delivering useful products, but the decade-long time frame for its projects gives researchers plenty of room to start research from scratch and investigate first principles before focusing on applications. “I really like the combination of realistic practice and the futuristic research and development there,” he says. “But I found I’m more interested in that futuristic development side. That’s why I decided to go into academia; it has more freedom and fewer boundaries. You can have bold ideas—even bolder than in industry.”
UW-Madison, Liu says, is a place with the resources and the colleagues that can enable him to pursue these bold ideas. His lab will be a full-stack operation, taking analog circuit ideas from design through fabrication, testing and integration into system applications. “I want to push the boundaries of high-frequency wireless integrated circuits and systems for ubiquitous applications,” he says.
Among those applications, he thinks his circuit designs could improve autonomous vehicles, smart glasses and other emerging technologies that use wireless networking. But he also sees his advances leading to new applications, like “body area networks” or sensors that collect and monitor health data, as well as devices that can receive high-speed satellite internet anywhere in the world, no bulky base stations needed.
Liu also wants to design a platform that allows radio-frequency circuits to take advantage of the full electromagnetic spectrum, instead of being locked into one narrow range. This could dramatically improve the energy efficiency of devices and better use the spectrum.
Besides his circuit- and system-level interests, his most immediate goal is to develop methodology level platforms that leverage artificial intelligence to improve design productivity. Analog circuits use a wide array of electrical components, including transistors, amplifiers, filters and regulators, each with advantages and drawbacks. The current design process is tedious, taking many researchers multiple years to bring a design from the drawing board into reality.
“In my future vision, I want to use AI to not only accelerate the design cycle,” Liu says, “but most importantly, to discover entirely new design spaces for novel RF structures that cannot be directly designed through human experience but can deliver far better performance.”