SILICON VALLEY

--the place of entrepreneurial legend. For decades famous and not so famous companies have been born here--in garages, on dining room tables, in cheap strip-mall office space.

Stanford University--long an incubator of future entrepreneurs, but a university without a formal curriculum in fostering start-up companies until only recently. Just last year Thomas H. Byers, an associate professor in Stanford's School of Engineering, launched the Technology Ventures Co-op program.

With that launch, the celebrated region and the renowned university became partners in a program that aims to give engineering students hands-on experience in a Silicon Valley start-up company.

The program, known as TVC, is offered to undergraduate or coterminal (simultaneously receiving bachelor's and master's degrees) engineering majors who have at least a senior standing. It starts in the spring term with a course on the management of technology ventures, including product and market strategy, financing, team building, and the challenges of managing growth. The program's centerpiece is a summer internship at a high-technology start-up, along with an evening class to help students process their experiences. It winds up with a fall debriefing class, during which students develop case studies and present them to the class--with their summer bosses in attendance. TVC interns are paid at the prevailing rate for entry-level engineers in Silicon Valley, typically $4000 per month. The companies are eager to hire these students because they are the "best of the best" at Stanford. What's more, the companies believe successful internships can help their overall recruiting at Stanford. And the TVC program has strong ties to the venture capital community, a link entrepreneurs are usually eager to nurture.

The program has two objectives. For one, Byers believes that undergraduate students can be taught to be entrepreneurs through a formal engineering program, instead of learning simply by doing what previous generations of entrepreneurs have done. By understanding how start-ups develop ideas and business plans, raise funding, and hire management and technical personnel, the students learn the ground rules of entrepreneurial success, something they can hope to apply to their own ventures someday.

A second program aim has a research component. By throwing students into the demanding and often turbulent waters that are an engineering start-up, and by asking them to keep daily diaries, the university obtains a rich dataset for understanding the relevance of engineering curricula in entrepreneurial settings in the real world. In fact, faculty-led research projects based on the TVC data have already begun, including "Newcomer Experiences in High Tech Start-Ups" [see the related "How start-ups motivate new engineers"] and "Venture Capital Funding Decisions." Byers hopes that this program will influence engineering education nationwide, leading to the creation of similar courses at other major engineering schools.

The first class of 10 TVC students interned during the summer of 1996. They were in high demand; most received internship offers from several companies. They had different goals--some wanted to learn marketing, some to gain command in business management, and some to get in-the-trenches design experience.

Six of the 10 were electrical engineering students. Five of those six offered to share their experiences--along with sections of their personal journals--with IEEE Spectrum. These diaries offer some glimpses behind the scenes at some of today's Silicon Valley start-ups.

Christopher Gori
Age: 23.
Born: Schenectady, N.Y.
BSEE: Stanford University, June 1996.
MSEE.: Stanford University, December 1996.
Technical focus: computer architecture and microprocessor design.

Previous work experience: At age 16, Chris worked in software development at Foxware Corp., Foxboro, Mass., a small software company designing turnkey systems for warehouse management. At age 19, he had had a summer job doing some hardware design, prototyping a motor control board, and developing control software at Micrion, Peabody, Mass. At age 20, he had spent the summer at Hewlett-Packard Co.'s medical products group, where he worked on control software for a test and measurement system.

By the summer of 1996, Chris was two classes short of his master's degree. He planned to take a class in logic design and in an introduction to computer networks. He had already completed his required undergraduate classes, including computer architecture, software engineering, and assembly language, and had finished a graduate level architecture class, in which he learned to use Verilog, a type of circuit design software.

Chris joined the TVC program because, based on his summer and part-time job experience, he believes he flourishes in the flexible start-up environment. He was interested in finding out more about the business aspects of starting a company. To sign up for TVC, he had declined possible internships working on microprocessor projects at Digital Equipment Corp. and Intel Corp.

Chris interned at Granite Systems Inc., Palo Alto, Calif. The company was incorporated in February 1995 by Andreas ("Andy") Bechtolscheim, co-founder of Sun Microsystems Inc., Mountain View, Calif., and David Cheriton, professor of computer science at Stanford University, to address the market for gigabit Ethernet products. Granite was creating standards-based multilayer gigabit Ethernet­switching technologies using powerful application specific integrated circuits. The company was sold to Cisco Systems Inc., San Jose, Calif., in late 1996 for $220 million. The company had about 20 employees when Chris started his internship in June 1996.

While Chris had interviewed with two other networking companies that were interested in hiring a TVC intern, he chose Granite because of his favorable impressions of Andy Bechtolsheim and Mark Ross, director of hardware.

"I was just amazed by the pace of the conversation with Andy," he said. "It was the proverbial 'drinking from a firehose' experience, and you don't get much lower water pressure from Mark, either. Both of them seemed to clearly want to go places."

Later Chris realized with pleasure that not only were they heading places, but they weren't going to trample their employees getting there, which he had heard was rare in a start-up.

Ross recalls offering the internship to Chris because of his background, attitude, and a recommendation from Tom Byers (TVC program director). Impressed with Chris's strong hardware and software skills, Ross said, "It is rare to find someone graduating with an EE degree who has as much software skill as Chris does. This is quite desirable because the process of modern ASIC design looks much more like software than it did 10 years ago. Strong software skills are a tremendous asset."

About a week before the official start of the internship, Chris had a one-hour meeting with Ross to talk about which project he would be assigned. Ross was well prepared. He handed Chris a sheet of paper describing his responsibilities, deliverables (including verification of one logic block and design of another), and what he hoped Chris would get out of the summer, including learning about ASIC design in a real-world setting and contributing to Granite's success.

Ross assigned Chris responsibility for two activities: verification of one logic block and the design of a reusable block for ASIC components. But, he said, because of Chris's "natural talent and interest, Chris championed investigation of PC platforms and performed some very important system administration roles" along with his logic design work.

"The tasks chosen for Chris were meant to challenge him and to provide useful work for the company," Ross told Spectrum. "We had so few people at the time Chris was hired that it was important to use him for something useful. Both assignments were designed to give Chris exposure to phases of the design process not normally found in a university environment. Verification and lab work are not normally emphasized in school."

On his first day at Granite, Chris arrived at 9:05 a.m., worried that he was late. "This turned out to be almost silly," he said, "since I would later find out that people were coming and going nearly all the time between 8 a.m. and midnight." (Chris ended up working typically from 10 a.m. to 7:30 p.m.) He met his office-mate, got his phone, computer, and e-mail set up, and filled out the standard Human Resources forms. Ross introduced him around the company, and by 3 p.m. he was at work. For his verification project, he was to report to the designer, while for his logic block design, he was to coordinate with a number of people. The specifications for both modules had already been written, which, Chris said, reduced the ambiguity of what he had to do.

Ross, who was responsible overall for Chris's work, expected regular progress updates every week or two; other than that, Chris operated on his own.

Chris found that he was readily accepted by the rest of the staff, 90 percent of whom were fellow engineers. With so much work to go around, the attitude, he said, was "the more the merrier."

Chris found his first assignment--verifying a section of Verilog code through simulation, then prototyping the module in a FPGA (field-programmable gate array) and testing it--challenging, even though, because of his class work in Verilog, he had thought he was prepared. In fact, Ross recalled, working in the laboratory was one of the few times Chris needed help. Logic synthesis and the company's logic analyzer were new to him. However, Ross said, "Chris's determination and problem-solving skills made him successful. Exposure to verification and lab work helped round out his skillset."

"The prototyping and part selection (of the FPGA, UART (universal asynchronous receiver/transmitter), EEPROM, and RS232 driver) was more difficult than I expected," Chris said. "There were several questions about how we would go about it, and about what parts should be tested first, since there were varying levels of schedule pressure. I also got to learn about a synthesis tool, which was one of the gaping holes in my prior experience."

To Chris's surprise, his high-school French proved useful; Granite has a large French-speaking contingent among its engineers.

The biggest problem, Chris said, "was the rapid change in some of the design requirements, so back-tracking ended up happening."

Chris found that expectations for him at Granite were high--higher, at least, than he had experienced during his work at Hewlett-Packard. At an established company, he said, "you weren't thinking on your feet as much, nor were you handling unanticipated problems like installing fast Ethernet hubs or setting up modem pools. In a start-up, because there is so much to be done, you have a much better chance to test the waters in a variety of areas. And you get better access to people involved directly with high-level decisions." Chris said he learned not to be afraid to take risks, as long as they are calculated risks, and he wishes he had been more aggressive in taking on responsibilities.

Chris completed his master's degree in December 1996 and Granite offered him a full-time job. He signed a contract in August 1996, complete with stock options, promising to return after he had completed his master's. The company was acquired in September, before Chris's re-turn. Granite is now part of Cisco's Workgroup Business Unit.

Stephanie Hannon
Age: 22.
Born: Reston, Va.
BS Computer Systems Engineering: Stanford University, March 1997.
MSEE: Stanford University, March 1997.
Technical focus: Networking, operating
systems, and distributed systems.

Previous work experience: Stephanie held summer internships at Intel Corp., Santa Clara, Calif., working in the Intel Architecture Lab on corporate demos that contrasted the performance of the Pentium with 486 microprocessors; participating in the Measurement, Architecture and Performance group on a central processing unit instruction-level performance evaluation of the Pentium versus Pentium Pro.; and co-authoring a circuit design methodology tutorial for the P7 project. At Stanford, she worked as a section leader, teaching small sections of introductory computer science classes. She also was involved in porting the Linux operating system to SimOS, a research project developing a machine simulation environment for studying uniprocessor and multiprocessor computer systems.

Stephanie attended a magnet high school for science and technology near Washington, D.C., and planned to become a biomedical engineer. She was accepted at both Stanford and the Massachusetts Institute of Technology; Stanford won the toss because of the appeal of the California climate. After taking--and being fascinated by--the introductory computer class required for all engineering undergraduates and the introductory electronics class, Stephanie changed her major to computer systems engineering.

She joined the TVC program in her senior year, shortly after being accepted into the MSEE program. She had taken a number of computer and electronics classes, including introductions to circuits, signals and systems, digital design, several logic design classes, computer architecture and design, object-oriented systems design, assembly language programming, operating systems, compilers, and Internet technologies. Classes in communications engineering, networking, and business management remained ahead of her.

After working at Intel for three summers, Stephanie had been looking for something different to do after her senior year. She had snagged a research fellowship, and thought a summer working with Ph.Ds. would help her decide whether or not to continue her education to that level. But then she saw a flyer advertising TVC, and was attracted to the idea of working in a small start-up company because her time at Intel had never given her the feeling that she was a critical part of a team.

"This is understandable when teams are hundreds of people and summer interns are only there for three months," she told Spectrum, but "I thought the demanding pace of a start-up would give me the opportunity to be an important, contributing member of the team."

"When I was accepted into TVC, I decided it was a once-in-a-lifetime opportunity and I gave up the research assistantship," she said.

Like Chris, Stephanie joined Granite Systems. She interviewed with the software team, which was primarily made up of Ph.D. students from Stanford, after being inspired by Bechtolsheim, who had lectured to the TVC class. "I thought he was a brilliant engineer as well as a savvy businessman and would be an excellent role model," she said.

She was impressed with Bechtolsheim's attempt to build a company out of what he considered the best and the brightest engineers from Stanford and Silicon Valley companies.

She discovered that a Stanford student who had always impressed her with his questions and inquisitiveness in class was working at Granite, and that most of the engineers she met had long lists of credentials, with many belonging to the "billion dollar club"--hardware engineers who had worked on projects that had shipped over a billion dollars of products. It was also clear to her, she said, that the technology they were working on was the "cutting edge" in hardware and software, pushing the C++ language in directions she had never seen by using a sophisticated simulation environment to debug software before hardware was available. She thought she could contribute to the group, because she was focusing on operating systems in her master's program, and because of her work with simulation research at Stanford. (And, she said, she also liked the free Snapple that was always on hand.)

Matt Zelesko, who was a senior software engineer at Granite when Stephanie was hired, recalled that Stephanie had a strong technical background through her course work as well as relevant experience through her research work in the SimOS group at Stanford. "Her performance at Stanford revealed both the ability to quickly develop and utilize new sets of skills and the ability to balance several projects simultaneously," Zelesko told Spectrum. "I regard these traits as prerequisite to a successful internship, particularly at a start-up."

"Furthermore," he said, "Stephanie was genuinely excited about Granite. There was an energy and enthusiasm always present at Granite, starting with Andy and David and carrying all the way through the engineering team. I believed that she would excel in our team of insanely energetic engineers."

On her first day, Zelesko showed her her new office. It was completely empty. He helped her round up a few tables and push them in, then unpack a computer and monitor. He then left her alone to get the computer set up and software installed. By the end of the day, she had two office mates--another summer intern and a Stanford researcher who would later join the Granite staff--and the room was filled. The company's growth was palpable.

The software team that she joined had eight people. She and Zelesko had come up with a list of potential projects for her to work on--for example, adding performance metrics into the operating system--but that list was discarded because a new network traffic generation tool from Netcom Systems, called Smartbits, had just arrived and Stephanie was asked to learn how to use it.

This task evolved into a project to build a test environment, employing the new equipment, that could be used by both quality assurance engineers and software engineers to validate functionality and test performance. The test project was to serve as a foundation for all of Granite's quality and performance evaluation and for the testing of hardware and software related to the switching platform. What was required of Stephanie was learning Granite's proprietary operating system and internal software structure, assisting with the development of prototype hardware, defining the interface, implementing it in code, and then providing documentation and sample uses.

Said Zelesko: "Stephanie was able to take some wild and crazy ideas dreamed up by the software team and turn them into reality. The particular requirements and goals for her project were only vaguely defined and even less documented. She was able to take ideas from other members of the team, define the goals, and independently design software interfaces and implementations to meet those goals."

Commented Stephanie: "One of the most exciting things was that the project was not clearly defined when I got into it. That gave me the opportunity to research what tools people had been using to create tests, what the limitations and capabilities of those tools were, and what their wish lists were for a new test environment. I really felt like I was designing a system with input from many engineers, not just implementing something that had been fleshed out before I ever got there. It was technically challenging because there were many components I had to get up to speed on very quickly. And I was designing a system that would be used by people with varying levels of programming skills, which required an easy to use but powerful interface."

Stephanie felt her education gave her a base to hold intelligent conversations with her co-workers, but the job itself gave her a wealth of knowledge about system design, simulation, programming, and networking.

The company's internal World Wide Web site also proved to be a key tool for Stephanie; all meeting minutes, design documents, and other material generated by Granite were posted, and she could often solve a problem by sifting through that information. If the Web failed her, she would approach her colleagues. Zelesko appreciated that Stephanie's questions to him and the other engineers had been well thought through before being posed. Besides informal discussions, he and Stephanie met weekly to review her accomplishments, evaluate goals, and discuss problems, as well as to review general company happenings, including product strategies. She had opportunities, too, to discuss the business model and market opportunity with Bechtolsheim.

Stephanie also found a mentor in Jo Beth Metzger, who oversaw quality assurance and adapter development. "Jo Beth had an open door policy," Stephanie said. "If I needed to talk to her about technical problems, business issues, or planning for the future, she was always available. She has spent most of her career in start-ups, and she was a wealth of information for me."

"I learned that communication skills are key to getting your job done," Stephanie told Spectrum. "At the beginning of the summer, I was intimidated by the people in my group and had a hard time asking for help. Eventually I realized that they didn't expect me to figure out everything on my own, and were willing to discuss design trade-offs and solutions."

Stephanie's project was critical to Granite's future, which was exactly the experience she had hoped for.

"Start-ups typically can't afford to assign toy projects to interns," Zelesko said, "and Granite was no exception." Zelesko was pleased with Stephanie's work, particularly her ability to evaluate business as well as technical tradeoffs. "On one occasion," he said, "Stephanie revised a portion of the implementation that, while not providing as many technical features, was easier to both implement and deploy."

In general, he believes, computer science curriculums lack an emphasis on iterative design and the design of large-scale software systems. They focus instead on writing toy programs from scratch that are used only for one assignment, and therefore miss the issues of designing maintainable code and programming within an existing large code base. "Undergraduate research opportunities at Stanford, such as the simOS project Stephanie was involved with, expose students to these issues while still in school," said Zelesko.

Stephanie was offered a full-time job at Granite on completion of her internship. She signed a contract promising to return after completing her master's. As is typical at start-ups, the contract included stock options. Shortly after she signed, Granite was acquired by Cisco Systems; her stock options are now worth in the six-figure range.

Is Stephanie disappointed to be working at a big company again? "The free Snapple is gone," she said, "but this is an opportunity for me to explore a different corporate culture and business model. If I don't like it, I can move on, but my experiences at Intel, Granite, and now Cisco are all valuable. They help me figure out how to pick companies I want to work for in the future and perhaps may help me decide how to run my own company someday."

Alison Hu
Age: 23.
Born: Cleveland, Ohio.
BSEE: Stanford University, June 1996.
MSEE: Stanford University, June 1998.
Technical focus: Circuit design.

Previous work experience: Alison spent three summers during college as a laboratory trainee and then as a technical associate at AT&T Bell Laboratories in Holmdel and Murray Hill, N.J. There, she designed microwave discrete and integrated circuits and developed guidelines for software programmers who were working on interfaces for networking applications, as part of a technical advisory committee. She also interned during her senior year of high school at NASA's Ames Research Center, Moffett Field, Calif., where she helped analyze data from biological experiments.

In June 1996, Alison had just received her BSEE and had already started course work for her MSEE (something students are able to do simultaneously with undergraduate work at Stanford). She had completed courses in analog design, semiconductor physics, and statistical signal processing; she was looking forward to digital design and radio frequency design courses the following year.

She joined the TVC program because she wanted a high-level overview of how a company operates, something she thought she'd get at a small company.

Alison interned at NeoMagic Corp., Santa Clara, Calif. NeoMagic makes embedded dynamic RAM (DRAM) multimedia accelerators for laptop computers using a proprietary low voltage silicon technology that integrates large DRAM memory and complex logic subsystems into a single chip. Its 3.3-V chips are found in laptops from Acer, Compaq, Dell, IBM, and Sharp, among others. The company was founded in the summer of 1993. When Alison came on board in 1996, the company had about 50 employees.

She interviewed with the vice president of engineering, senior design engineers, applications engineers, the vice president of marketing, and the chief financial officer. Her first impression of the company was that it was a bit disorganized; it was in the process of moving to a bigger building, boxes were everywhere, and people were hurried. But she was intrigued enough to go back because she felt the company had a very compelling technology and she liked the people. On her third visit, she talked to some people whose excitement about what they were doing was catching, who convinced her that there were valuable projects she could complete in one summer, and who offered to mentor her.

Said Tom Rogge, a senior member of the technical staff for NeoMagic: "The company was looking for someone who was friendly, adaptable, and could work well with the other people in the company. We had to find someone who wanted to participate in a relatively unstructured environment, and someone with a very strong educational background who could jump right in and be productive without too much initial training." Alison fit the bill.

Unlike many EEs, Alison wanted a position in marketing, and NeoMagic offered one. She was told that she would have the opportunity to work on projects with other groups.

On Alison's first day, Rogge, assigned as her mentor, took her around the company and introduced her to several engineers. At first, she said, some people in the company were not sure how to work with a summer intern, but once people understood the TVC program, they were enthusiastic and helped her as much as they could. Administrative assistants showed her how to work the voice mail system; the facilities supervisor showed her how to work the building's alarm system. Her mentor and his boss, Niall Bartlett, the vice president of marketing, presented her with suggestions for projects, but before she could get started, a more critical task emerged because the company's engineers needed some crucial data, and Alison was off and running.

Completing assignments for this task involved performing timing tests on NeoMagic's hardware, running graphics benchmark tests on Neomagic's chips, and collecting the test data, a task that fell to the technical marketing department. Rogge had been running the tests for the first assignment before Alison started, so they were all set up. But for Alison to run them, she had to learn how to use new tools, including the Microsoft Developer Studio and Soft Ice (from Nu Mega) quickly. Since Rogge was out of the office traveling most of the first few weeks, she had to put forth some effort into figuring out when and how to run the needed tests.

While her assembly language programming background helped her understand some of the code that drove the tests, Alison found that since she had no training in graphics, she had to hit the books to determine what the tests for the second assignment measured. In her college classes, she often had to learn how to use new tools, so learning yet another set of tools was easy for her. When she didn't understand the code well enough to run the tests, she would go for help to software engineers, who she said answered her questions readily. She also spent a fair amount of time talking to the hardware engineers at the company to find out what they wanted tested.

Rogge said that Alison worked on challenging tasks that needed doing, because in a small fast-growing company, it is essential that everyone contributes. "She was able to fit into the small company atmosphere," he told Spectrum,. "She was very self-reliant, and was prepared to go and work with other people to resolve problems. This showed a very mature attitude for someone just completing their degree."

She also, Rogge said, tried to instill some culture into the engineers by dragging them off to concerts at Stanford.

Alison liked her experience with a start-up company because of the lack of bureaucracy and the pervasive attitude that, if you need something done, you do it yourself, because there isn't anyone else who will do it. She enjoyed interacting with different departments, feeling comfortable going to anyone in the company for an answer to her problem.

She found being young, Asian, and female all worked to her advantage. Because she was young, she felt, people were slow and patient with explanations. Being Asian gave her a bond to the company's engineering workforce, which is largely of Asian background. Being female, she felt, didn't hurt her as an engineer, and gave her a way to bond with other female employees in other departments. This bonding gave her a better picture of the company as a whole and helped her sort out which things that she might do or might avoid if, one day, she runs her own company.

She did make mistakes, she said. For one, she didn't ask as many questions as she should have in the beginning. "If I had been more aggressive, I could have learned more," she told Spectrum. She also made the mistake of assuming that the code that people gave her was correct, and wasted time trying to figure out why something didn't work when she wasn't given the right thing to begin with.

Said Alison: "If you invest the time and energy to work in an intense, start-up environment, you will learn a lot more in less time than you would at a larger company where the need for you to learn quickly is not necessarily as crucial."

Benjamin Jun
Age: 23.
Born: Cleveland, Ohio.
BSEE: Stanford University, June 1996.
MSEE: Stanford University, March 1998.
Technical focus: Control systems.

Previous work experience: While attending a math/science magnet school in Silver Spring, Md., Ben spent three summers working at a defense contractor, the Center for Computing Sciences, in Bowie, Md. There he did a simulation of an experimental massively parallel processor, wrote kernel and diagnostic software for early versions of the machine, and developed runtime-optimized code for a bank of the processors to be installed in a special version of the Cray-3/SSS supercomputers. During college, he interned in the San Francisco office of Bain & Company, a Boston-based management consulting firm. While an undergraduate at Stanford, he worked with a group that was investigating low-cost methods of magnetic resonance imaging and with another that was creating a satellite-navigated autonomous farm tractor, guided by the Global Positioning System.

As an engineering undergraduate, Ben's favorite classes were those in the laboratory that let him see EE principles come to life and the nontechnical ones that gave him the overall "big picture"--like ethics, sociology, and industrial seminars. When he joined TVC for the summer between his BSEE graduation and the beginning of his master's studies, Ben had been wanting to work for a small company for some time. "I think this allure runs in the blood of Stanford students who live close enough to Silicon Valley to taste it," he told Spectrum.

The previous summer, he had approached several start-up companies on his own, hoping to receive an internship, but struck out. "Companies would be interested in my skills, but would tell me that I would be of little use to them if I couldn't arrive to work on Monday and begin coding in earnest the same day," he said. So he jumped at the chance to join a program that included a formal internship at a start-up and a chance to learn about business, instead of just doing technical tasks.

Ben first read about the firm Ideo Product Development, Palo Alto, Calif., in West Magazine, a Sunday supplement to the San Jose Mercury News. He was so impressed with Ideo that he asked himself, How could I not try and work for this company? By strict definition, Ideo is not a start-up. The company, which does product development for other firms (its designs include computers, toothbrushes, cellular phones, ski goggles, and the mechanical whales that starred in the movie Free Willy), has 300 employees in Palo Alto and around the country. It started as David Kelley Designs 18 years ago.

Assigned to an Ideo engineering team that was working for a venture-backed start-up company, Audible Inc., Wayne, N.J., Ben was able to experience the start-up environment through regular interactions with that company's founder and chief executive officer, Timothy Mott, who was a Xerox Parc veteran and cofounder of Electronic Arts and Macromedia, and with Guy Story, Audible's vice president of engineering.

Ben's interview at Ideo, he said, was probably the most unusual interview he ever had. First, he spent an hour with Tim Taylor, Ideo's head of electrical and software engineering, discussing company cultures and Ideo's processes before going over Ben's background. Then came the office tour, which, he said, "involved passing through creative office partitions, like a DC3 airline wing, and walking under beach umbrellas, a vintage WWII searchlight, and nearly 20 suspended bicycles that people had ridden to work that morning. All around me was the evidence of playful minds and a willingness to try many times: the office was littered with the gutted remains of dozens of cellular telephones, boxes of used prototypes, and inspirational products that the engineers thought looked, performed, or were designed well. I went home hooked and hoping for an offer."

Ideo offered Ben the internship, Taylor told Spectrum, because he had an interesting resume with broad experience--from supercomputing software programming and ruggedized system design to management consulting and hardware design of massively parallel systems. Also, Ben was clearly very capable and thoughtful, and demonstrated the curiosity required to do creative product development.

His first day started with the company's weekly "all hands" Monday morning meeting, which gave him a glimpse of various projects. There was little employee orientation; Ben slowly discovered for himself who to ask questions of, where to pick up his printouts, and how to use the company's information resources. He met Scott Brenneman, who was to be his mentor and project leader on a team that was just beginning the process of designing a hand-held electronic communications device.

This product was a new concept, the Audible System, to be introduced in late 1997, consisting of a palm-sized audio processor that played material such as "spoken books" through a headset or through a car radio simply by laying it on the car seat or dashboard. Material was obtained by downloading it to a PC from the Internet and then transferring it to the audio processor.

Ben recalls that when he found out that he would be working on a project for the founder of Electronic Arts, "I had to keep pinching myself. In junior high, my computer interests were fueled by a bunch of Electronic Arts magazines and programmer interviews; I hoped to someday become proficient enough to work there."

His first assignment was to help in component selection for Audible's product.

Said Ben: "It was the first time I needed to carefully scrutinize the design issues of power consumption, features, and cost. Navigating through a bewildering array of options of digital signal processors and other chips was difficult; I gradually learned how to look for components with a 'designer's eye' and learned how to catalog information about a part for future reference."

Ben moved into the board-level design of the product, and went on to develop specifications for content delivery and work on technology licensing. His internship stretched into the fall and he postponed his graduate studies for three months. After he returned to school, he continued as a consultant to the company.

Ideo's Taylor told Spectrum that Ben required little guidance once he started on a task; it was up to him to know when to ask for assistance. "For example," Taylor said, "when developing an intellectual property position and investigating technology licenses, Ben was pointed in the general direction, and from then on worked directly with the client and attorneys to complete the task." Once a week, Ben and Taylor would get together to discuss his project as well as the company as a whole. "On a daily basis, he was part of the project team, which is self-managed--in other words, he wasn't managed," said Taylor.

Ben had a lot of access to the Audible executive team, since no one else at Audible was, at that point, involved in the project.

"Preparing technology briefings and presenting design decisions to the Audible team helped me gain the perspective of being in a start-up's trenches," Ben told Spectrum. "It was an engineering environment unlike any I had previously experienced. Product decisions are made with a clean sheet of the product vision ahead of you, the reality of resource constraints at your back, and with the constant urgency of the time-to-market clock."

Because there was so much to do--both at Ideo and at the start-up company for which the Ideo team was working--Ben felt little emphasis was placed on his relative inexperience or age, and that the breadth of the product issues being considered meant that no question was too silly or unimportant.

Ben advises engineers considering a job at a start-up to ask a lot about the company and make judgments on its soundness--its goals, investment partners, and financial condition. He also urges a student intern or new engineer to find an engineering mentor to help swim "in the understaffed and unknown environment that most start-ups live in."

Diana Yan Fu
Age: 24.
Born: Nanjing, China.
BSEE: Stanford University, June 1996.
MSEE: Stanford University, June 1997.
Technical focus: Digital signal processing, communications systems.

Previous work experience: As a technical associate at AT&T Bell Laboratories in Murray Hill, N.J., during the summer of 1995, Diana designed and conducted experiments in electron backscatter, aided the making of a next-generation electron beam lithography tool, and wrote simulation and data analysis programs. At Stanford, she was the residential computing coordinator, managing computer and networking resources in five undergraduate dorms. She was also a research assistant for a magnetic resonance imaging project and for a physics research project requiring her to design IC masks and build driver circuits for experiment. In 1992 Diana worked on Stanford's Solar Car Project, developing electronic components.

When Diana joined the TVC program, intrigued by the opportunity to learn about business management, she had completed courses in signal processing, electronics, and IC and analog circuit design. But even though her advanced communications courses were ahead of her, she was interested in the field of communications. So when the opportunity arose to intern at Spectrum Wireless, Mountain View, Calif., founded in February 1995 to develop major components of the base station subsystem infrastructure for GSM (Global System for Mobile Communications) wireless personal communications networks, she jumped at the chance. Spectrum Wireless is a subsidiary of Cellular Telecom Ltd.

"I chose Spectrum Wireless because the job assignment was interesting: I was going to join a two-person team, working on a network design application for the Spectrum Wireless network." Diana said. "Andy Walters, the project leader, had just started the architectural level design when I joined him. Participating in the design and implementation of a new software project of this scope was a rare opportunity to learn many aspects of software engineering."

According to Karen Coates, vice president at Spectrum Wireless, the company selected Diana because of what it viewed as her long-term potential. "Diana is not only very strong technically and quite articulate, but she has an understanding beyond her years of organizations and people," Coates said. "She is also highly self-motivated. She will be an influential player in a corporation in the future."

On Diana's first day at work in June 1996, the company was one and a half years old, with 35 employees. Its office space was a Mountain View warehouse, without even cubicle partitions. Andy Walters, her supervisor and the project leader, showed her around, introduced her to the engineers who, despite the fact that they were working on different projects, would be her nearest neighbors in the building, and handed her some books to read on software architecture and visual C++. Her first assignment was to set up a series of measurement units and conversion factors to be used as the software library. This assignment was a small part of one of the company's key projects.

"This was a bit of a gamble," Coates said, "but we felt confident enough in her abilities to go with it."

"I did not feel my education adequately prepared me for this challenge," Diana said, "since only one of the computer science courses I took taught any C++. I was an EE major, after all." Learning to use Visual C++ was a challenge for her, but a co-worker on the project was helpful, as was the software's on-line tutorial and manuals.

She worked closely throughout with Walters. "Because of the nature of the application," she recalled, "I had to learn a lot about the architecture of the network as a whole, which provided many opportunities to work with engineers from other groups."

Diana did a variety of tasks. "I don't recall her sweeping the floor, but I'm sure she would have if asked," Coates said. "We make it clear to everyone that in a start-up, you do whatever needs to be done."

At the end of the summer, Coates asked Diana's team leader if he would like Diana to stay on the project as a half-time employee while she completed her master's program, or replace her with a full-time engineer. Without hesitation, Coates reports, Walters recommended that Diana stay on.

For books offering insights on entrepreneurship and management techniques, see Entrepreneurship, Management, and the Structure of Payoffs by William J. Baumol (MIT Press, Cambridge, Mass., 1993), High-tech Ventures: The Guide for Entrepreneurial Success by C. Gordon Bell and John E. McNamara (Addison-Wesley, Reading, Mass., 1991), and Recapturing the Spirit of Enterprise by George F. Gilder (ICS Press, San Francisco, 1992.

Stanford University's Technology Ventures Co-op Program has its own World Wide Web page: http://www.stanford.edu/group/stvp/. It describes how interested employers and students can participate in the program.

The Web site of NeoMagic Corp., Santa Clara, Calif. (http://www.neomagic.com/) lists career opportunities with the company.

Cisco Systems Inc., Santa Clara, Calif., which acquired Granite Systems Inc., Palo Alto, last year, has an "Employment Opportunities" page. For more information, see http://www.cisco.com/jobs/.