By Lawrence M. Fisher  |  Photos By Bart Nagel

Political and corporate leaders of a certain age often speak of the United States' diminished status in scientific and engineering leadership, and they can quote a variety of scary statistics to support their concern. Look out 10 years and it’s easy to spin scenarios in which Bangalore, Shanghai and Singapore have taken the technology lead from Boston, Sunnyvale and Seattle.

But wander around the Berkeley campus for a few days and that dire vision just doesn't fit the reality on the ground. There you'l find a bunch of energized brainiacs working on projects that can only be described in technical language, like: ”way cool.“

And Berkeley Engineering is not alone. With federal funding for basic research in decline and corporate R&D budgets slashed in favor of short-term earnings gains, institutions of higher learning are increasingly called upon to foster innovation and entrepreneurship.  At the same time, the problems that demand engineering solutions have changed, and engineering education has evolved along with them. Engineering was only recently a very vertical vocation, with aeronautical engineers drawing airplane wings, civil engineers designing bridges and electrical engineers laying out circuit topology. But today's engineering applications are broadly multidisciplinary, and any new product—like   Boeing's  forthcoming 787 Dreamliner—draws as much upon the talents of software savants and materials mavens as aerospace adepts.

Furthermore, the information technology revolution has inserted computing into every engineered system. From automobiles to medical equipment to microwave ovens, engineered products increasingly rely on microprocessors and embedded software. The creators of those products need not only fluency in traditional engineering disciplines like thermodynamics and high-level programming languages, but also a fundamental understanding of the contexts in which these systems operate. The masters of these hitherto disparate disciplines must all learn to speak a common language and work in globally distributed project teams. They must be deep and broad and adaptable.

The good news is that those adjectives describe many of the students you might meet at Berkeley Engineering. They’ve grown up reading Dilbert and have no intention of spending their careers toiling at thankless tasks and taking abuse from a pointy-haired boss. They also grew up hearing Steve Jobs say that technology can change the world, and they are anxious to make contributions of their own.

Incubating entrepreneurs
Consider Sonesh Surana, a graduate student researcher in electrical engineering and computer sciences (EECS), whose dissertation focused on an inexpensive, long-distance WiFi technology that he deployed in rural southern India. The technology supports a telemedicine system that now serves 100,000 patients in remote areas with no doctors.

Doing that work confronted Surana with the acute power problems faced by rural areas and led directly to QV Sense, the startup he is incubating with Jason Stauth, another EECS Ph.D. candidate who specializes in efficient power management. QV Sense, which earned them a 2008 award from Berkeley’s Center for Entrepreneurship and Technology (CET), is developing a system to make solar power generation more efficient.

The CET was initiated in 2004 to teach entrepreneurial skills to engineering students in addition to their technical training. Through its Global Venture Lab, the CET sponsors competitions and global networks to support the students in developing new technology ideas. It is “a great resource for first-time entrepreneurs,” Surana says. “We save on cost, which is a big deal for a startup. But more important, we have a weekly entrepreneur lecture series, after which the entrepreneur or venture capitalist is available to give us feedback. Often the feedback is pretty sobering.”

That kind of exposure gives students a reality check and a sense of the competitive landscape, Stauth says. The team is now in the process of reducing their design from a circuit board to a single chip, using the skills they learned in their engineering courses. “The fact that we’re also learning state-of-the-art engineering is putting us ahead of our competition in the area of solar power management,” Stauth adds.

The Global Venture Lab, buried within Bechtel Engineering Center, looks like the unused classroom that it is except for a few pieces of electronic testing equipment from the likes of Hewlett- Packard and Tektronix. But it has one great feature for startup companies: it is rent free. Together with the prize money of $10,000, that is a great boon for a bootstrapped enterprise.

What other industry leaders are saying about educating engineers

There is a distinction to be made between educating students to be entrepreneurs and to be entrepreneurial, says Yogen Dalal, a managing director at the Mayfield Fund venture capital firm in Menlo Park, California. “I don’t know that any institution can train people to be entrepreneurs, any more than you can teach a kid to be an athlete. But they can teach you how to be entrepreneurial and how you can team up with entrepreneurs even if you are not one yourself,” he says.

Classes at Berkeley Engineering do not stress entrepreneurship, which is taught in dedicated classes through the CET and Haas School of Business. The CET’s charter is to give engineers and scientists the skills to lead and innovate and to commercialize technology in the global economy.

“Our first premise is that meaningful problems are the seeds of opportunities,” says Ikhlaq Sidhu, CET director and professor of industrial engineering and operations research. “Engineering education has changed. There was a time when it was all theory, fundamental theorems and proofs. That is critical, but at some point people recognized the value of laboratories. It’s not enough to have a theory; you need to try things. The next logical step is, can you do these things in the constraint of the real world?”

Is the United States falling behind?
Despite the allure of startup success and the inclinations of venture capitalists, not every budding engineer wants to launch a startup, and not every bright idea warrants a new enterprise. Often, the inventors of a new technology are not the first to recognize its best application. The original microprocessors were developed at Intel and Texas Instruments to power digital watches and calculators, not the personal computer, which grew out of hobbyists using the new chips in home-brewed projects. And many of today’s most pressing problems—from energy independence to personalized health care—are too multifaceted to be addressed by startups, which tend to focus on a single technology.

Indeed, many innovations behind today’s most important technologies did not come from entrepreneurial startups. The transistor was developed at Bell Labs, the research arm of the original AT&T; the graphic user interface for computers was developed at Xerox PARC; and the Internet emerged from ARPANET, the federally funded computer network designed to facilitate communication between government labs.

But Bell Labs is no more, Xerox PARC is much diminished from its glory days, and today’s companies are too pressured delivering quarterly gains to shareholders to invest much time or capital in basic research that may never provide a return. While researchers are cheered by President Obama’s inclusion in the stimulus package of $20 billion for basic science, they worry about its priority among so many pressing economic problems and wonder how far it can go in recovering 30 years of government spending cuts in research not directly tied to defense.

In her 2008 book Closing the Innovation Gap, serial entrepreneur Judy Estrin argues that Silicon Valley is riding on the investments and inventions of the past. Our society, Estrin writes, has so devalued the scientific professions that today’s students are more inclined to look elsewhere for their career choices.

“The leaders of our academic institutions,” Estrin says, “need to be rethinking their programs. I’m not advocating revolutionary change all at once but efforts in all the major engineering schools to think long term about how the departments are organized and how they work with other schools in the university.”

Others argue that the decline in the United States’ technological lead may be more perceived than real. A Duke University study examining statistics showing vastly more engineering graduates in India and China than in the United States found that those Asian countries had included in the count low-level technical positions—what Berkeley Engineering dean Shankar Sastry calls “commoditized technologists”—and that on an apples-to-apples basis, there was in fact relative parity among the three nations, in spite of vastly larger populations in the east.

Nevertheless, many engineering tasks are fungible and, in a market-based global economy, these are already moving to lower-cost countries. The growing numbers of engineers in countries with lower labor costs means that the United States and other developed nations must aim their graduates at positions that emphasize innovation and leadership. This conclusion is evidenced by the plethora of books that, like Estrin’s, advocate a renewed emphasis on innovation; enter the word innovation on an Amazon.com book search and you get nearly 300,000 titles.

An ecosystem of innovation
Berkeley Engineering’s innovation agenda emphasizes a host of newer research centers that, like CET’s Global Venture Lab, are built around the dual goals of pushing innovative technologies to be rapidly embraced by industry while simultaneously introducing engineering students to real-world experience. They range from the four-campus Center for Information Technology Research in the Interest of Society (CITRIS)—which emphasizes collaboration with researchers in law, business, economics and public policy to speed integration and adoption of new technologies—to the more focused Berkeley Wireless Research Center, which is working on designs to support next-generation wireless communications.

These centers rely heavily for support not on federal grants but on industry partners with an interest in the work and in the student talent pool. The Parallel Computing Laboratory, or Par Lab, opened in 2008 under the direction of EECS professor David Patterson. The space it occupies in Soda Hall was designed to deemphasize old-school hierarchies like private faculty offices, instead using open space, low cubicle walls, easy chairs and white boards to inspire spontaneous and constant creative collaboration among faculty, students and industry reps.

The Par Lab was funded exclusively by Intel and Microsoft through a national competition to explore the future of parallel processing, a form of computing in which multiple problems are calculated on multiple processors simultaneously. As increasing a single processor’s speed has run into the limitations of heat dissipation, semiconductor manufacturers have switched to placing multiple processors on a single chip to achieve higher speeds. Now the greatest challenge facing computing is making it easy to write programs that execute efficiently in parallel on systems with multiple processors per chip. The work is inherently multidisciplinary.

“Rather than classic computer science where you build this thing and then get somebody to write an application for it, we’re bringing in multiple domain experts who have applications that are just thirsty for processors,” Patterson says. Historically, the most processor-hungry applications were in defense, oil exploration or materials science, but today they are just as likely to be in an imaging system that can provide real-time views of internal organs or a video game with movie-like, user-created characters. And in the near future, they could be in your very own personal electronics.

“The Par Lab vision is for parallel computing to be accessible to everyone,” says Ph.D. student Sarah Bird, who came to Berkeley specifically to work in the lab. “We are working on mobile applications to help the average user with things like image recognition—where you could hold up your camera phone to the Eiffel Tower; it recognizes that you’re in Paris and gives you a list of places to go—and speech recognition, where your laptop or cell phone could write out digital text for everything said in a meeting.”

Although Bird’s specialty is computer architecture, or hardware, she is now working with multidisciplinary teams—including researchers from Lawrence Berkeley National Laboratories, Intel and Berkeley’s music and math departments—on what she calls “the whole stack,” including hardware, operating systems and applications. “I was drawn to working with other people who do all these cool things in areas that I am not expert in.”

But how do you teach a student to come up with new ideas for such applications? You have to create the right ecosystem, says EECS professor and Par Lab researcher Kurt Keutzer.

“Everybody has potential for innovation, but for a high degree of it you need a pool of gifted students like we have here at Berkeley,” says Keutzer, who was chief technology officer at Synopsys and a researcher at Bell Labs before joining the engineering faculty. It takes four things to make innovation happen, he adds: the best resources, including people; a collaborative environment that stimulates generation of ideas; incentives in the form of recognition or remuneration; and real problems to solve. Like the CET’s Sidhu, Keutzer emphasizes the importance of finding the right problem. “At the core of innovation is basic human curiosity and some real need,” he adds. “Necessity really is the mother of invention. You have to expose people to real problems.”

The problem solving extends even beyond venture creation and entrepreneurship, says Jan Rabaey, codirector of the Berkeley Wireless Research Center (BWRC), which also uses an industry-funded collaborative model, bringing together faculty, graduate students and industries. BWRC researchers talk about a not-too-distant age when every individual might have thousands of wireless devices, from heart-rate monitors to home air quality sensors, all networked, all running all the time.

“Our research can have a broad impact, influencing government regulations and public policy, even creating new industry directions and trends,” Rabaey says. “This is just as important as getting a new idea and venturing it out to create a new company; we need societal leaders and that is part of our training process.” The research being done today at BWRC, for example, might change the way wireless is regulated 15 years from now or introduce an innovative wireless technology approach in the developing world, Rabaey adds. And, like many Berkeley labs, BWRC has chosen to forgo proprietary claims to its inventions.

“Everything we do is public domain and available to everyone,” Rabaey says. “We say, you want to take it, go for it.” This model is in the tradition of open source software, the basis for the Linux operating system, in which multiple developers publish and freely share their work. It suits the university’s need to publish its findings, providing the researchers a kind of real-world peer review. It also speeds the adoption of new ideas by industry and benefits society by giving companies a free platform upon which to base their own innovations.

Open source has been given even broader application through the concept of open innovation, as described by Haas School of Business professor Henry Chesbrough. Although it differs from open source in the exclusivity of licensing, open innovation suggests that, in a world of widely distributed knowledge, companies cannot afford to rely entirely on their own research but should instead buy or license processes or inventions from a network of suppliers.

Broad new world
Today’s emphasis on multidisciplinary project teams means that engineers have to leave the cubicle if they are to play a critical role. They need to know what they know, but they also need to know what they don’t know. They need a key understanding of the piece of a project they work on directly, but also an ability to pattern match, to understand what other team members are talking about, what they need, and who they can turn to for help.

While acquiring that degree of depth and breadth means that tomorrow’s engineers may have to work harder than their forebears, it is unrealistic to expect a new generation of polymath men and Renaissance women, says Alfred Spector, vice president of research at Google. “It’s not worthwhile to tell our engineering students that they have to do the impossible,” he says. “They don’t have to be superhuman. People will operate within their discipline at a great level of detail and then have to work with others on issues of multidisciplinary integration.”

Implicit in Spector’s comment is the realization that engineers are going to need better social skills as well as technical smarts to succeed in a multidisciplinary environment. Indeed, Daniel Goleman, author of Emotional Intelligence, has consulted with several high-tech companies on how to improve their engineers’ interactions with peers. In Silicon Valley, where new hires are screened primarily on the basis of cognitive ability, many companies end up with a lot of brilliant engineers without social skills, Goleman says. “It’s a problem in businesses, particularly in the tech sector, where the culture of type A achievement and system thinking would intrinsically reward high-functioning Aspergers,” he says, referring to the autistic condition characterized by difficulty with social and communication skills.

Beyond acquiring a basic proficiency and ability to communicate in related disciplines, the engineers of tomorrow will need to be cognizant of entirely different fields of science, particularly biology, says Paul Saffo, a futurist and consultant who teaches a class at Stanford called The Future of Engineering.

“For an engineer 50 years ago, the symbol was a wrench and a mechanical pencil; 10 years ago, a computer and a CAD system. Ten years from now, it’s going to increasingly be a set of biological principles,” Saffo says. The never-ending drive toward smaller and more complex devices will favor molecular solutions over manufactured ones, and biotechnology has already demonstrated that one-celled organisms like yeast can be programmed to produce very intricate structures in vast quantities. Tomorrow’s computing devices might well be built of proteins and peptides rather than silicon, Saffo says.

Back in the Par Lab, David Patterson and Kurt Keutzer don’t seem worried about Berkeley’s ability to feed the system with engineering talent in the coming decade and beyond. “There may be countries where engineers are more highly prized than they are here in the United States,” Keutzer says, “but I tend to think that people who have the native talent will naturally be drawn into the science, engineering and technology fields.”

Is there a problem with innovation? “It’s hard to see it being a problem from the research side,” Patterson adds. “We think we’ve just scratched the surface of uses for information technology.”