(FORTUNE Magazine) – Back in the 1980s the U.S. was seen as a rusting industrial giant. Better get ready, said the doomsayers, for a future of selling pizzas and movies to the Japanese. Since then American companies have been on a sustained restructuring and productivity ride. Times are still not easy; the country has lost something like three million manufacturing jobs since 1998. But people like Leo Gerard, Joe Houldin, and Stuart Parkin have kept the factories humming with American innovation.

JOE HOULDIN

The man whose passion is manufacturing.

There wasn't much competition in 1988 when Joe Houldin applied for the director's job at the Delaware Valley Industrial Resource Center (DVIRC), a startup with the mission to help southeastern Pennsylvania's small and medium-sized manufacturers succeed. "Most people thought there was no future in manufacturing," recalls Houldin, 53. He disagreed, and has brought his clients an array of services, including quality management systems, financial analysis, strategic planning, marketing, and website design. His team has taught more than 5,000 factory workers the wonders of just-in-time techniques and Toyota-style manufacturing.

Funded by public and private sources, the DVIRC has grown from a one-man, two-desk shop to a 40-strong operation with $9 million in annual revenues. Its efforts in 2004 generated $69 million in value-added dollars, according to the U.S. National Institute of Standards and Technology. It is "one of our highest-performing centers," confirms Kevin Carr, director of the federal Manufacturing Extension Partnership (MEP), whose mission is to help manufacturers modernize. The investment pays off: Every dollar the state allocates to its seven industrial resource centers (IRCs) generates $15 in taxes, says Tom Palisin, Pennsylvania's manufacturing ombudsman. He counts the Philadelphia-based DVIRC among his top three performers: "The results are very good."

One case in point: Rod Bayard, president of Bestweld, which makes pipefittings for aircraft carriers and is based in Pottstown, Pa., credits DVIRC coaches with teaching Bestweld's 52 workers techniques that cut its delivery time from 11 weeks to five. Bestweld's improvement satisfied key customers like Newport News and attracted new ones. (The DVIRC also helped train my own small factory to implement continuous quality-improvement techniques.)

The creation of the IRCs was a response to the hammering that Pennsylvania's manufacturing sector took in the 1980s. With big factories closing down, much of what was left of the manufacturing base was small, niche-oriented--and struggling. "We were getting our clocks cleaned," says RoseAnn Rosenthal, at the time head of the finance division of the Philadelphia Industrial Development Corp. (PIDC).

That's where Houldin came in. Raised in a working-class Irish-Catholic neighborhood in Philadelphia, Houldin watched the daily choreography of men leaving for the local General Electric or Westinghouse plants. He was struck by their dignity of purpose. After Houldin's father suffered a severe illness that left him debilitated for years, his co-workers stepped in: "They took care of me and my mom."

Among the few in his neighborhood to attend college, Houldin supported himself with factory work until he graduated from Villanova in 1979. He taught in a private school while getting a master's degree in city and regional planning from Catholic University, then joined the PIDC. Working firsthand with the area's small manufacturers, he observed that "a lot were poorly run." The DVIRC offered the chance to do something about that.

It was, in retrospect, a natural fit: Houldin has never been short on confidence--nor has he faltered in his belief in the value of manufacturing. He is a passionate advocate, convinced that vibrant communities stem from production work. "The city's poverty and blight are inextricably bound to its loss of manufacturing," he contends, leaning his hefty 6-foot-1 frame across a round meeting table. "Drive through North Philadelphia," an area characterized by falling population and rising despair, "and you'll see what happens when you don't get it right."

The DVIRC's own development had a few hitches. "We spent the first year developing and testing concepts like information services that it turned out nobody wanted," recalls Tony Girifalco, executive vice president. "They had more fundamental issues, such as getting product out the door and worrying about inventory and payables." The agency adapted. Indeed, Houldin takes pride in operating it with a mix of idealism and hard-nosed business judgment. Fee-based income, which comes from services like consulting or product development, has grown 34% a year since 1996, reaching $2.6 million. Houldin has ambitious plans--more services, more income, more clients. "It's going to be a blast," he concludes.

STUART PARKIN

The man who is making disk drives disappear.

Stuart Parkin, 49 and lean as a greyhound, talks fast, walks fast, and drives fast, racing his BMW over the hilly campus where he toils for IBM. While few people have heard of him, most American households have at least one product that is a direct result of his work.

Framed awards crowd Parkin's walls at IBM's arts-and-crafts--style Almaden Research Center in San Jose. "Stuart's one of the most extraordinary scientists in the world," says his boss, Tom Theis, director of physical sciences at IBM Research. Parkin's work with thin-film magnetic structures enabled the first commercial use of spintronics-based technology. Spintronics relies not on an electron's charge but on its "spin," or magnetic power. Like an infinitesimal dervish, each electron generates its own magnetic field. But until the early '90s, electrons were thought to be too chaotic to harness. Parkin and others saw potential. By applying a magnetic field, they can organize the "spin," or "current," of the electrons, some 2,000 times smaller than an atom, to store and transmit data. The result is machines that are faster, cooler, and smarter than ever before.

Raised in Manchester, England, Parkin was always competitive and studied hard, excelling at mathematics. "If you didn't let Stuart use his brain enough, he'd be awake," his mother recalls. He had his pick of colleges, selecting Trinity at Cambridge, he says, "because that was where Isaac Newton went." He got his Ph.D. in 1980 and landed at IBM as a post-doctoral fellow in 1982. Invited to remain a year later, he found he fit in well with a company that prides itself on inventing the products it sells.

Parkin's first big breakthrough came in 1991 when he set out to improve the capacity of IBM's hard-disk-drive line. He succeeded by finding a metallic material that optimized the performance of giant magnetoresistance (GMR) structures--atoms-thick layers of metals with the capacity to store huge amounts of data. The result was to increase disk-storage densities 20-fold. Used in every disk drive today, Parkin's work in spintronics led to iPods, digital cameras, Smart Cards, and capacious laptops. Theis allows, "We made a lot of money on that." Last year IBM sold its hard-disk-drive business for $2.1 billion to Hitachi Data Systems.

"I want to change the world," says Parkin, who was named an IBM fellow, the company's highest honor, in 1999. Most recently he's been after the semiconductor industry's Holy Grail, attempting to formulate a spintronics chip with the data density of DRAM, the speed of SRAM, and the nonvolatility of Flash. To that end Parkin has transformed his "magical" magnetic circuit structures into nonvolatile magnetic memory devices (MRAM).

Parkin's hands and feet jiggle as he describes merging spintronics with conventional computer chips: "It's beautiful technology. We can leverage this huge silicon-based industry and just incorporate MRAM on top of that," he says. The advantages are formidable. Spin chips, or MRAM, will hold memory magnetically, much the way the earth, one giant spin device as it were, maintains polarity. Combined with conventional computer chips, spin-magnetized layers store data and conserve power because they don't require a charge. No more sizzling cellphones and laptops. No need to save work or wait for computers to reboot. No need to recharge your electronics for days, or maybe ever. "IBM has a lot of know-how to make MRAM feasible," observes Jim Handy, an industry analyst for Semico. "With this, you're getting into the realm of science fiction."

IBM has delayed its introduction of MRAM in order to integrate it with microprocessor chips. Meantime, Parkin is researching a way to make MRAM more sensitive--and thus to make hard-disk drives obsolete. Literally unable to sit still, he appears ready to burst at the prospect of the spin-driven universe ahead. Bounding around one of his several laboratories, where scientists tend giant machines, he is concocting new molecular electronics, "magnetic racetracks," and "spin transistors." One day the results of his work may allow the rest of us to function with something like Parkin's alacrity.

LEO GERARD

The union man who saved steel.

"Not all union leaders have devil's horns," says Wilbur Ross, founder of International Steel Group (ISG) in Cleveland. "Steel would probably be half the size it is today without Leo." Ross refers to Leo Gerard, 57, president of the United Steelworkers (USW). By applying sound economic principles to two of the nation's most troubled industries--steel and tires--Gerard is helping to save jobs and preserve the country's industrial base while strengthening his union at the same time.

In 2002 the steel industry was in dire straits, with 35 companies having gone belly-up since 1997. The problem, says Gerard, was simple: "Too many companies were trying to sell and make steel." He sought a modern-day Andrew Carnegie to buy the better companies, combine them, and thus create a stronger industry. Ross, a financier known for buying companies in trouble, saw opportunity. In 2002 he created ISG out of LTV, Bethlehem Steel, and Acme Metals. Around the same time, U.S. Steel bought National Steel Corp., and Allegheny Technologies acquired J&L Specialty Steel. The USW played a key part in each deal.

This consolidation has made the American steel industry significantly more efficient, profitable, and productive--albeit with 50,000 fewer workers than in 2000. The USW's deal with Ross, for example, locked in investment to modernize production, allowed flextime, and established an industry-backed voluntary employee beneficiary association, pooling profits to provide health care to retirees. For his part, Ross sold the corporate jets and the golf courses and cut back corporate staff to the bare bones--about 70 people. Other steel companies followed suit. Those moves, combined with the lucky confluence of China's surging demand for steel, antidumping tariffs, and the falling dollar positioned Ross to cash in. In 2004 he parlayed his $400 million investment in ISG into a $4.5 billion sale to Mittal Steel of Rotterdam, the world's largest steel company.

"We changed the world through collective bargaining," exults Gerard. There were tradeoffs, of course, in the form of lost jobs, higher medical copayments, and outsourcing of nonsteel tasks. But the industry is in better shape than it has been in years. U.S. Steel's stock price has risen sixfold since April 2003, and the company turned in a profit last year of almost $1.1. billion.

Gerard played a similar role when Goodyear wanted to shift a chunk of its U.S. operations offshore in 2002. "We said, 'Bullshit,'" snaps Gerard. After protracted negotiation, Goodyear kept all but two of its 14 U.S. plants, trading manufacturing and technology upgrades for productivity gains and a 16% reduction in its 19,000 USW employees. Gerard also displayed a deft business touch in Canada. In 1991, as a union official in Ontario, he had the idea of the union and its workers buying the troubled Algoma steel mill. With help from the Canadian government, they did--and turned it around. Union members are now minority shareholders; ton for ton, Algoma remains the most profitable steel company in North America.

In demeanor Gerard is like the economics professor he once wanted to be: rumpled, articulate, and analytical. His background, though, is deep blue-collar. He was raised in Ontario in the small company town of Lively, where his father worked as a coal miner and union organizer. He would listen on the basement steps as his father and other men "plotted about things like how to get safety glasses--I thought it was cool." As a teen, he passed out union leaflets and picketed. He later tried to juggle night school with a day job at the local smelter, but his growing involvement with the union eventually forced a choice. He chose the union movement, working his way up the ranks before becoming president in early 2001.

Like his corporate counterparts, Gerard has done some consolidation of his own and has embarked on a merger program with other North American unions in such disparate industries as tire and rubber, industrial wood, and flint glass. He also wants to develop more international links. At a time when union membership has declined to only one in every 12 American workers in the private sector, Gerard is energized and committed. "There is no free and open democracy in the world that doesn't have healthy, independent labor unions," he says.

FEEDBACK fortunemail_letters@fortunemail.com. Stories from the Industrial Management & Technology section can be found at fortune.com/imt. Executives in manufacturing and research and others eligible to receive FORTUNE's Industrial Edition can subscribe by calling 888-394-5472.