Advanced Computing in the Age of AI | Thursday, March 28, 2024

Addressing the Digital Manufacturing Labor Shortage 

Gerry Ledford offers some ideas on how to increase the supply of skilled labor in the first of a two-part series.

Part I: Increasing the Labor Supply

A critical constraint on greater use of advanced digital manufacturing technology is the shortage of skilled workers who can run high performance computing (HPC) software and hardware. In this first of two blogs, I will offer some ideas on how to increase the supply of skilled labor. Part II focuses on what individual companies can do to compete in the labor market and get at least their fair share of the labor that is available.

The place to start in increasing the labor supply is raising the investment of colleges and universities in HPC. The skill level required for HPC is high enough that almost all employees using these skills will need at least a bachelor's degree in a technical field. Despite a national unemployment rate of almost 10 percent, including a shockingly high unemployment rate even among technical professionals, colleges and universities are not producing enough qualified graduates in a field that is growing rapidly and is desperately short of qualified candidates. How can this be?

First, colleges and universities are not proficient at adapting to the needs of the job market. Courses and programs change slowly in academia. Indeed, institutions of higher education typically do not offer programs that anticipate emerging needs of the labor market. They are much better at traveling well-worn paths and turning out graduates who are able to perform in well-established fields.

HPC faces another problem. The types of professionals needed to master HPC have skills that lie at the intersection of several disciplines, including engineering, mathematics, information technology, and manufacturing. I spent 16 years on the faculty of the University of Southern California's Marshall School of Business, and I learned that developing interdisciplinary programs and coursework does not come easily to academia — despite considerable lip service about the need for such programs. The academic reward system in general does not promote interdisciplinary work. For example, a young faculty member seeking tenure may see interdisciplinary work as risky compared to working in established disciplines that are well recognized by his or her department and mentors.

Given this set of conditions, what can HPC users do promote the production of graduates who can help them now? Three strategies can work.

First, companies can encourage colleges and universities to produce the types of graduates that they need. University administrators (for example, Deans of engineering schools) tend to be more responsive than faculty members are to the needs of the business community. Deans value the financial support that companies and business leaders potentially represent, and the employability of the school’s graduates is an important issue for administrators. Although Deans do not have the level of control over faculty that company executives have over their subordinates, Deans have powers to influence, cajole, and reward that should not be underestimated.

Most Deans would be extremely impressed if representatives of even a relatively small group of companies — say five to ten — were to meet with him, advocate a common position, and outline the skills that are highly valued and in short supply in the labor market. An exchange of views about what would be needed to encourage training of more HPC students may be fruitful for both sides. Even more impressive to Deans would be a promise of financial support for HPC training. Relatively modest amounts — say contributions of $5,000 to $10,000 per participating company — may make a large difference in purchasing needed hardware and software, creating formal interdisciplinary programs, providing scholarships that build student interest, and making relevant faculty research possible. Ultimately, identifying established faculty members willing to champion HPC is critical. Some faculty may be drawn to HPC by the consulting opportunities, others by research opportunities. The coin of the realm for faculty in research universities is academic journal publication, and providing opportunities for faculty members to gain access to companies for research on HPC may be extremely helpful in recruiting faculty champions. Part of the discussion between interested faculty members and company representatives can be about opportunities and limitations on both sides.

Second, the HPC field may be a terrific area in which to create a cooperative education (co-op) program, which allows students to alternate between periods of academic and corporate work. Co-op programs have a long history at some engineering schools, but are rare in many HPC disciplines. Companies interested in HPC need not convert the entire school to the co-op model — they merely need create a pool of HPC co-op students. There are clear advantages on both sides. The participating companies may get a tremendous recruiting advantage through the co-op experience, they may help students understand exactly the kinds of skills that they need to build, and they may obtain skilled, albeit junior, employees at affordable wage levels. Students get excellent financial support for higher education at a time when sources of financial aid are shrinking and tuition costs are rising. Students also get real-life exposure to the challenges of the field, and they greatly increase their chances of employment post-degree.

Finally, employers may encourage HPC experts to develop or expand consultancies. This allows companies to stretch their use of scarce talent, leveraging the skills of senior HPC experts to multiple companies. Using consultants may be a viable alternative when there are not enough experts to go around. Consultancies also can help overcome the problems that arise from the scarcity of HPC talent, especially in small- and medium-sized firms. Senior consultants can provide the mentorship, technical guidance, and other assistance to junior consultants that is difficult for companies without established HPC programs to offer. In a sense, this strategy may allow companies to outsource the recruiting and development of HPC talent to a consulting firm.

We have reviewed three strategies for increasing the availability of employees with HPC skills, all of which are relevant to large as well as small companies. These strategies require collaboration across organizations. In Part II, I will consider how an individual firm can compete in the labor market to hire and retain the HPC talent it needs.

EnterpriseAI