The disparity in talent demand among companies is becoming more pronounced.

On one hand, positions in front-end development, testing, operations, marketing, and design are facing elimination. On the other hand, there is a high demand for roles such as FDE (Frontier Deployment Engineer), Agent developers, AI evaluators and security specialists, and full-stack engineers.

This shift is not merely a localized adjustment by a single company but represents a broader restructuring of talent demand across the entire internet industry. A clear indicator of this restructuring is evident on the recruitment platform Maimai, where the talent pool is now categorized into two distinct groups: AI talent and non-AI talent.

Screenshots circulating online from Maimai's backend showcase an "AI Talent Label" prominently featured in the filtering menu and beside candidate avatars.

Talent Acquisition: Early Campus Recruitment and Changing Logic

In previous years, internet autumn recruitment was concentrated in "Golden September and Silver October," a well-established industry norm. However, since last year, this timeline has been repeatedly moved forward.

Baidu initiated its campus recruitment in July, Alibaba and ByteDance opened their 2027 graduate recruitment channels this week, and DeepSeek commenced large-scale recruitment as early as June. Some top AI positions are now offering "unlimited salaries," and ByteDance has even established a separate early-bird application channel for AI product managers, with the quickest route to an offer taking just two days.

Why the urgency? Because AI talent typically takes one to two years to become integral to business operations, and early recruitment is crucial for building a competitive talent advantage.

However, this year's talent acquisition strategy differs significantly from last year's.

Previously, AI talent acquisition primarily focused on securing top algorithm engineers—master's and doctoral graduates from prestigious universities like Tsinghua, Peking, Fudan, and Jiao Tong, first authors of top conference papers, and those more proficient in TensorFlow/PyTorch than in-house frameworks. These individuals were scarce, highly sought after, and commanded starting salaries in the millions.

But this year, the real demand lies elsewhere.

Alibaba's current recruitment encompasses eight major categories, including algorithms, R&D, chips, and products; ByteDance's campus recruitment for AI product positions has surged by 20%; and Meituan's campus recruitment targets core positions in technology, products, and operations.

What companies are most in need of now are individuals capable of applying AI in practical business scenarios—those who can adjust Agents, build Workflows, and integrate RAG into real systems to deliver tangible results. These individuals may not have published top conference papers, but they possess the ability to integrate models into enterprises, create tools usable by non-technical staff, halve operational processes within two weeks using Coding Agents, reduce customer service costs by a third, or triple designers' output speed with an AI tool.

As the platform with the highest concentration of AI talent, Maimai was the first to recognize this structural shift. Maimai CEO Lin Fan once shared a set of data: 80% of AI talent is active on Maimai, with a large number of HR and headhunters actively seeking out AI talent on the platform. Executives from companies like DeepSeek and Zhipu personally recruit, with profile accessories reading "Resume Wanted" and signatures stating "Long-term Recruitment" and "Email Address."

This time, Maimai's separate labeling of AI talent underscores that frontline recruitment service providers have also acknowledged the changing dynamics of corporate talent acquisition.

Elimination: Optimization Continues, AI Becomes the "Scissors"

While there is a frenzy for certain talents, layoffs are occurring elsewhere.

Domestic internet giants have been continuously optimizing their workforces in recent years; one can search "Big Factory + Layoffs" on Maimai or Xiaohongshu for more details.

The situation is equally challenging overseas: According to Morgan Stanley, the U.S. white-collar service industry experienced a 15% increase in unemployment due to AI last year.

The tech internet industry fared even worse, with Meta laying off 8,000 employees, Amazon 30,000, and Microsoft 8,500. In May 2026, 40% of U.S. layoffs were attributed to artificial intelligence.

This is the "AI Employment Gap"—a restructuring of job roles. The economy is experiencing a K-shaped divergence, and the job market is facing a similar fate. Next year, tech companies may truly only recruit AI talent.

Reshaping: External Frenzy, Internal Transformation

How are big companies advancing AI implementation? They're taking a two-pronged approach.

Externally, they're scrambling for "technical foundations" and "advanced application developers"—individuals with prestigious educational backgrounds (at least 985/211), papers, experience, and results. These individuals require long-term cultivation, and companies are locking them in early through campus recruitment.

Internally, transformation occurs in two ways: First, existing R&D engineers undergo AI training to transition into application development; second, a large number of non-technical staff directly use Coding Agents to transform business operations, proving their worth through efficiency gains.

Alibaba and Tencent have already initiated company-wide AI transformations. A Meituan campus recruitment official revealed that in 2026, a new "AI Competency Assessment Module" was added, with assessments required for technical, product, and operational roles. 70.12% of e-commerce companies have initiated or are evaluating salary adjustments linked to AI efficiency.

Industry demand has shifted from "basic algorithm R&D" to "implementation and application." A recruitment official bluntly stated, "The rarest are composite talents who can understand business and translate AI capabilities into industrial value."

So, where does this leave us ordinary folks?

The Only Way Out for the White-Collar Population: Being Labeled as AI Talent

To identify opportunities, observe who companies are recruiting. Refer to how Maimai, a recruitment platform with extensive interactions with tech internet companies, categorizes AI talent labels:

For the vast majority of white-collar workers, the first two categories may be too ambitious; AI business efficiency talent is the only viable path for breakthrough.

Listing a plethora of tools like Doubao and DeepSeek on your resume is futile; what matters is the results you've achieved with AI—increased work output by X%, shortened project cycles by X days, the traffic generated by your work, and the practical problems solved.

Merely listing tool names on your resume is insufficient; focus on showcasing tangible results.

Maimai CEO Lin Fan holds a radical view: True proficiency with Coding Agents means allowing them to run tasks automatically for over an hour. You can gauge your own level against this benchmark.

Achieving results is not enough; you need to be visible. Update your results and refine your resume or Maimai profile to ensure algorithms label you as "AI Talent," making it easier for HR and headhunters to discover you.

Acting Now Is the Best Time

OpenAI's internal AI penetration rate soared from 10% to 99.8% in just eight months. Domestic tech companies are closely following suit.

With the accelerated penetration of Agents, tech companies may truly cease recruiting non-AI talent next year.

But this presents an opportunity—under the new rules, the window for overtaking on curves has not yet closed.

Don't wait until the label is slapped on you to realize you've been relegated to the "non-AI" column.