The American labor market is frequently described as facing a talent shortage. Employers say they cannot find enough qualified workers. Workers say they apply widely and hear very little back. Public discussion tends to treat these two claims as opposing views of the same economy. In practice, they often reflect the same problem from opposite sides.
That contradiction has become harder to ignore in fields such as logistics, healthcare, technology, and advanced manufacturing, where demand for labor remains strong and hiring still moves unevenly. As of early 2026, the United States continues to report millions of open roles alongside persistent underemployment. Candidates who appear capable do not make it through screening. Employers invest in new recruiting tools and still report weak matches, high turnover, and long ramp-up periods after hire.
For Yunan Weng, the problem starts with how hiring systems read people.
“Many employers still operate with a ‘sorting first’ model,” Weng explains. “They reduce a candidate to a few signals, then make exclusion the first step. That might help with volume, but it leads to poor matching. You produce a system that gets better at rejecting applicants, rather than one that accurately identifies them.”
Weng, Chief Human Resources Officer at Yanwen Express, has spent much of her career working at the point of friction between labor demands and hiring design. In prior roles across high-growth and cross-border organizations, she saw a recurring pattern: companies had plenty of data, but their hiring and workforce decisions remained sprawling and fragmented across various platforms and tooling.
“Workers don’t experience hiring and engagement and on-the-job training as separate things,” she says. “A mismatch at the hiring stage will show up later in performance, as manager load, and eventually as turnover. Companies often absorb those costs downstream instead of addressing the source.”
The Matching Problem in the Labor Market
That view has shaped Weng’s reading of the current U.S. labor market. In her account, a large share of the difficulty comes from translation failure. “Hiring has a language problem,” Weng explains. “Employers describe roles one way, candidates describe themselves another, and the system can’t reconcile the two. Once that gap enters the process, the technology built around it often reinforces the mismatch instead of fixing it.”
Employers post roles with job architectures and credential filters. Candidates tell their story through prior titles, uneven work histories, adjacent skills, and are limited to local opportunities. The two sides may be closer than they appear, but the systems in between are often built to reward standardization rather than interpretation.
This is one reason the labor market can appear tight and inefficient at the same time. Employers are not always failing to attract talent. In many cases, they are failing to identify it accurately. For Weng, that poses as a structural problem; a fixable one.
Putting Together the HR Puzzle
Weng argues that the weakness of many hiring systems cannot be understood in isolation from the rest of HR. Recruitment, onboarding, training, performance, and retention are still managed too often as separate functions, each producing its own data and its own decisions. “A company hasn’t solved the hiring problem if it brings people in efficiently but places them into poorly matched roles once they arrive,” Weng says.
Her earlier work gave her room to test that problem across a diverse range of industries. At Born to Learn, she worked on AI-based screening tools meant to evaluate candidates through broader patterns of skills and experience rather than static keyword matching. At UPMC China, she helped integrate AI into HR systems to support workforce analysis, employee development, and more informed staffing decisions. At Yanwen Express, her work expanded into a larger operating context where recruitment, retention, forecasting, and internal mobility had direct implications for a growing logistics network.
This operational model aligns with findings from Weng’s research in the Journal of Economics & Business Management, which demonstrates that continuous feedback systems and predictive analytics significantly improve retention and employee development when applied within integrated HR environments.
Those experiences now form the background to a new U.S.-focused technology and HR project Weng is developing around the same question that runs through her earlier work: how can employers build workforce systems that identify talent more accurately, support people more effectively after hire, and adapt to changing labor demand over time? She describes the effort as a response to a persistent gap in the American job market between available workers, employer needs, and the systems used to connect them.
“Not every labor issue is a pipeline issue,” she continues. “Sometimes it can be. Sometimes employers do need more trained workers. But there are plenty more cases where the labor exists and the matching process doesn’t suffice. I’m interested in that middle layer, where we can better design the HR systems to improve who gets seen and vetted.”
Building A Better Way
Weng says she is interested in building systems that also support onboarding, adaptive training, retention analysis, and workforce planning in a more connected framework. The premise is that hiring decisions improve when they are informed by what happens after hire rather than treated as a standalone event.
Her thinking also includes a geographic dimension. Employment opportunities in the United States remain concentrated in major hiring hubs, while many workers in other regions remain outside the strongest recruiting networks or skill-building pathways. For Weng, that limits employers and workers at the same time.
“Geography shapes opportunities more than many employers admit,” Weng says. “We might have gone remote, but a person’s location often stands in for capability or readiness, sometimes perceived long-term value, even when the evidence for that is weak. If companies want broader access to talent, they have to widen their nets.”
That view places her within a broader debate in American HR. Companies have added more tools to recruiting and people operations, yet many workers still describe the hiring process as opaque, repetitive, and poorly aligned with actual skill. Employers, for their part, continue to report missed targets, costly attrition, and trouble finding candidates who fit the role as it is actually performed rather than as it was written on paper. Studies indicate that replacing an employee can cost up to twice their annual salary, highlighting the importance of getting hiring and retention right.
“When people talk about infrastructure, they usually mean transportation or digital systems,” she says. “Workforce systems matter in a similar way. They influence how quickly firms can respond and how fairly opportunity is distributed. And sometimes, how much productive capacity is lost between application and actual contribution.”
Her new project is taking shape inside that argument. Weng believes that hiring accuracy, employee development, and retention should be treated as a package deal, and many hiring and workforce problems become easier to address when tackled together.
“AI is already an important part of HR,” she says. Recent workforce analyses indicate that over 70 percent of organizations are incorporating AI into HR processes to improve hiring efficiency and workforce planning. “The harder question is whether it can support better judgment at the points that matter most, and whether it can help people move into the right work with a stronger chance of staying and growing there.”
As employers continue to search for ways to make hiring more accurate and workforce planning more stable, that question is likely to draw increasing attention. Weng’s project grows out of the view that the labor market’s difficulty is not only about how many workers are available. It is also about whether the systems designed to recognize them are doing their job.
