The Battle for Talent in Hard Tech: TSP Headhunters Respond with Technology and Resources

Enterprises cannot find the right people, and job seekers cannot find jobs - this is a phenomenon that continues to occur in the talent market in 2024.

Take the highly popular artificial intelligence industry this year as an example. Major companies and startups in this field are eager for talents in AI and robotics. Not only did ByteDance launch the Top Seed program for doctoral students, but more than 80% of Baidu's 2025 campus recruitment positions are related to AI technology.

However, even so, both employers and job seekers indicate that recruitment in the AI industry is not easy. On the job-seeking side, an algorithm master's student with internship experience in a leading large-scale model company informed 36Kr that they have repeatedly applied for positions in major companies but without success. In the social recruitment, many AI companies that are highly sought after in the primary investment market stated that their companies not only have difficulty recruiting algorithm talents, but also find it challenging to find suitable candidates for positions such as growth, operation, going global, and ecological business development.

The mismatch between the expectations of recruiters and job seekers is the cause of this phenomenon. This imbalance between supply and demand also puts tremendous pressure on headhunters, who serve as the bridge between the two.

In fact, starting from the past two years, many headhunting companies have needed to continuously face challenges. Multiple headhunters told 36Kr that nowadays, the talent pools of many large enterprises are more complete, and the candidates recommended by headhunters may overlap with those in the pool, resulting in no added value. Additionally, enterprises have become more dependent on their internal HR departments and have raised higher delivery requirements for headhunters.

The mismatch in supply and demand in industries like AI and robotics can be considered a "happy problem" - in more industries, such as real estate, there are no signs of expansion, and the demand for headhunters has almost dropped to the freezing point. In May of this year, Tongdao Liepin (06100.HK), a comprehensive headhunting service platform, disclosed its Q1 2024 performance. The financial report shows that Tongdao Liepin's revenue in Q1 was 471 million yuan, a year-on-year decrease of 6.4%, and the net profit was -13.2 million yuan. In 2023, some securities companies also pointed out that the headhunting business of Career International is also "under pressure".

Some soul-searching questions have emerged in this industry, which has a 30-year history in China: What kind of headhunters can adapt to today's market environment? What kind of headhunting companies do clients really need?

A New Industrial Opportunity Requires Different Headhunting Companies

Headhunters have almost grown along with China's new economy.

The story began in 1994 when the Australian recruitment service group Seek entered China and founded a headhunting company called "Zhilian". This is one of the earliest headhunting companies in China, and the subsequent "Zhilian Recruitment" is its Internet version.

Later, traditional recruitment websites declined, and the Internet + headhunting model began to be valued. In 2011, Dai Kebin, the Greater China Brand Manager of Procter & Gamble, founded "Liepin.com". Since then, the headhunter position has become more accessible to job seekers in various industries.

However, around 2015, the space for headhunting companies began to be compressed.

One important reason is that "Boss Zhipin", which focuses on quickly connecting the two ends of the industrial chain, uses AI technology to quickly match supply and demand, expanding its customer base and seizing the territory of headhunters.

Up until 2024, "Boss Zhipin" had achieved a revenue of 3.6 billion yuan in the first half of the year, with a profit of 670 million yuan. In contrast, the revenue of "Liepin", a representative of headhunting companies, was 1 billion yuan in the first half of 2024, with a profit of 45 million yuan. Many more headhunting companies have found themselves in a situation with no business to do.

The solution to the problem may also be hidden in history.

The year 2015, when "Boss Zhipin" began to rise rapidly, is very similar to the year 2024 we are in now.

In 2015, it happened to be the climax stage of the rapid development of the Internet. At that time, companies like BAT, ByteDance, Meituan, and Didi were in an upward trend. Coupled with the rise of models such as the sharing economy, subsidy competitions and talent wars among peers were everywhere. Nine years later, in 2024, the Internet industry began to contract, and hard technology industries such as computing power, robotics, artificial intelligence, and low-altitude economy have replaced the Internet as the new trend.

This industrial transformation has brought new opportunities to the headhunting industry - the professional threshold in the hard technology field is extremely high, and talents are in great shortage. Headhunters, a role that creates service value by opening up resource and information gaps, are once again needed by the market. Just this year, many founders of hard technology companies in fields such as robotics, artificial intelligence, new energy, and chips told 36Kr that they feel the intense industry competition and it is not easy to find suitable talents, and they need the help of headhunters.

These founders also hope that the headhunters serving them not only are familiar with the high-end laboratories in the industry but also can reach the senior talents scattered in the academia, industry, and research fields at the first time - only in this way can the headhunter role play a helpful role in the highly competitive industry.

This outlines the portrait of the new generation of headhunting companies in the market: In the era of hard technology, headhunters not only need to continuously exert their resource advantages but also quickly gain a deep understanding of the high-tech industry. Moreover, they also need to be good at using digital means to improve the overall delivery speed.

The Breakthrough Solution of Gathering Technology and Resources

It is not an easy task for many traditional headhunting companies that are accustomed to "relying on resources" to integrate scarce market resources and cutting-edge technologies in the field to tackle new industries that are unfamiliar and have high thresholds.

Especially in the current severe market environment, many headhunting companies are even less motivated to change themselves.

"It is very difficult to get business this year, and many headhunters are already preparing to change careers." A technology headhunter who has been in the industry for 15 years told 36Kr that his peers either jump to the client side or leave the recruitment industry completely.

But what they don't know is that in such an environment, there are still some headhunting companies that have entered new industries and achieved growth.

"We have completed a performance of several million yuan in the robotics field this year." Bruce Deng, the founder of the "Talent Shop Partner" headhunting company, told 36Kr.

The advantage areas of this headhunting company are semiconductors, new energy, Internet technology manufacturing and consumption. It has no accumulation in AI/robotics. It was only in the second quarter of this year that it started from scratch.

In the highly competitive robotics industry, this is almost a miracle.

Last year, robot companies such as Zhiyuan Robotics, Yinhe General, and Xingdong Era were established in batches, and many experienced headhunting companies have flocked into the field, directly starting a "fierce competition".

By this year, more robot/embodied intelligence companies have emerged. Many of the founders of these companies are industry experts. To recruit talents for them who want to facilitate smooth financing and bring their own traffic, the competition will only become more intense.

Bruce told us that in the face of the difficult and high-threshold robotics field, TSP's chances of success as a newcomer were not high at the beginning.

However, they chose to use a specially customized AI tool to improve the learning efficiency of headhunters and help analyze resumes, and finally won the first batch of robot clients.

Using the new generation of AI technology to analyze job descriptions is one of the important reasons for TSP headhunters to increase the success rate in the new field.

Previously, the resume matching technology in the recruitment industry basically adopted keyword matching. Although this technology is direct in effect, it is not flexible enough and often matches candidates who do not quite meet the requirements.

But this year, when the job descriptions of clients are input by TSP headhunters into an AI product debugged based on a large model, the powerful natural language understanding ability of the large model can directly include candidates who were originally beyond the matching range in the potential talent pool.

Bruce introduced that in the talent search process, TSP headhunters have become accustomed to the "combination of human and machine".

"We will combine AI and manual search methods. The results searched by these two modes may have a 50% to 60% overlap, but there will still be many unexpected candidates."

The "parent company" of TSP, the domestic leading information consulting company "Liudu Zhinang", is the one that provides AI tool support behind the scenes.

In 2018, during the project execution, Liudu Zhinang found that many clients, after receiving information consultation, would hope to invite experts to join the company due to their recognition of the experts' professionalism. Therefore, in order to further close the loop of the clients' needs, "Liudu Zhinang" established a dedicated headhunting department, which is the predecessor of TSP.

In 2022, with the increase in such client needs, TSP officially became independent.

At this stage, it not only directly established its own brand but also set up offices in Hangzhou, Shenzhen, Quanzhou, Shanghai, Xiamen, Fuzhou, and Suzhou.

By 2023, the independently operating TSP has achieved overall profitability.

TSP, hatched from Liudu Zhinang, still maintains a close connection with the former and "inherits" its digital gene.

In the early days of Liudu Zhinang's establishment, its CEO Li Shengxiang stated externally that the company would accumulate a reliable expert resource pool through big data and artificial intelligence technologies.

When executing consulting projects, Liudu Zhinang will automatically match the expert list to improve the fit rate of the needs. Li Shengxiang once introduced externally that for the execution of consulting projects, this combination of artificial intelligence and human labor saves about 80% of the manpower and 75% of the time compared to the traditional model.

In the era of large models, Liudu Zhinang established a technical team of more than ten people in Hangzhou, which will debug AI tools for both its internal business and TSP. TSP will also provide timely feedback to help the technical team improve the AI.

"At the beginning, the results retrieved by the system had a weak correlation with our needs." Bruce said.

In order to improve the effectiveness of the AI system, TSP requires each of its headhunters to accurately list the information points according to education, position, and job content, and split the resume information into modules, and then enter these data into the system.

The technical team of Liudu Zhinang will combine these data with Liudu's expert consultation reports and "feed" them to the model for training.

At the beginning, headhunters would enter 120 to 200 resumes per month. Subsequently, as the training effect improves, the proportion of resume entry will also be reduced.

Moreover, this collaborative model not only improves the model level but also enhances the resource efficiency of Liudu Zhinang and TSP.

36Kr learned that although the talent resource pools of Liudu Zhinang and TSP are now divided into two sets, they are still interconnected and can be synchronized and updated at any time. Currently, this pool has stored 5 million expert resources from all over the world.

Bruce further told 36Kr that TSP's core advantage areas are the FAB of semiconductors, the photovoltaic energy storage of new energy, the robots in artificial intelligence, the Internet of Things terminals, as well as the beauty, food, new retail, and catering in the consumer goods sector.

These areas are relatively popular directions in recent years, and they are not blue oceans for headhunting companies.

But the differentiation of TSP is that traditional headhunting companies mostly rely on various online platforms for talent search, and are limited to finding talents at the BCD levels. Relying on the empowerment of Liudu Zhinang, TSP not only covers all the BCD-level talents in the market but also includes the A-level talents that are difficult for traditional headhunting companies to reach.

Furthermore, since Liudu Zhinang continuously interacts with high-end talents through information consultation services, when the TSP team conducts the headhunting business hatched from Liudu, it can also quickly build trust with these high-end talents.

"Another difference between us and our peers is that our users have a high degree of stickiness." Bruce said that this long-term trust accumulated from Liudu Zhinang's information consultation business not only helps TSP quickly match talents but also enables it to obtain real and reliable background check information for clients.

Everyone Carries Technological Expertise Every Day to More Places with Concentrated Hard Technology