Quantitative private equity firms are all restless, and have set up AI labs one after another.

Developing the global capital revaluation of Chinese technology investment through DeepSeek, the foresight behind it is the layout of artificial intelligence in the quantitative private equity industry. Recently, the talent recruitment for AI Lab, the intelligence learning laboratory under the billion-dollar quantitative Kuande, and the newly launched AI Lab of the billion-dollar quantitative private equity Meng Xi Investment have attracted attention. They openly recruited a team of machine learning interns and shouted the slogan "become the factor to change the world." The development of quantitative private equity is accompanied by machine learning and deep learning technologies and is a group that successfully applied artificial intelligence technology to financial markets relatively early. In addition to companies like Jiukun Investment and Mingcheng Investment, which have labs and North American investment research centers, quantitative private equity firms such as Heiyi Asset and Pansong Asset, although not establishing dedicated AI labs, have AI applications integrated into their investment processes. In daily investment research, quantitative investment requires the application of deep learning to discover factors and an earlier exposure to artificial intelligence, leading to a deeper understanding. At that time, in order to meet investment needs, a large number of chips were invested in, combined with a group of smart minds, lacking neither money nor manpower. This is also the basis for cultivating purely technical large models in the quantitative field. DeepSeek defines itself as a small company, with the values of making technology better and the goal of AGI (Artificial General Intelligence). By promoting open source, it drives the progress of the technology community. Therefore, there are no short-term considerations for commercialization or financing. In the industry's view, without the commercialization pressure brought by financing, quantitative private equity teams are usually not large in size. Efficiency brought by agile teams is what helps AI companies with a quantitative background go further. As these billion-dollar quantitative private equity firms continue to join, the development of artificial intelligence in China seems more promising. It is worth noting that many companies have stated that layout of AI does not necessarily mean entering the big model space, as there are still many innovative ways to implement at the application level. "DeepSeek's large model is already very advanced and open source, so the company has no need to create another large model," one top quantitative private equity firm said. Exploring cutting-edge technology, Meng Xi Investment establishes AI Lab "The future is already here, waiting for you to unbox." After Meng Xi Investment's AI Lab was established, it opened its intern recruitment process. According to the recruitment notice, the main recruitment target is AI researchers in machine learning, with three main responsibilities: developing quantitative trading strategies using machine learning models; tracking cutting-edge models and technologies in the machine learning field and trying to apply them to quantitative finance; and researching, analyzing, and statistically studying historical data using machine learning and deep learning methods to find relevant trends and patterns. Financial media reporters also learned that although it is an internship position, interns can access live factor libraries and anonymized platform data. The company hopes that these young people can play a supportive role while providing creativity and diverse thinking. Similar to other quantitative companies recruiting technical staff, the AI Lab's educational requirements are complemented by candidates who have won various competitions such as Kaggle. Interestingly, the AI Lab is located in Hefei, while the office is in Shanghai. Why did they choose to recruit in Hefei instead of leveraging Shanghai's university resources, going the extra miles to recruit in Hefei? It is understood that this may be related to the founder of Meng Xi Investment, Li Xiang. Public information shows that Li Xiang is from Anhui and graduated from the University of Science and Technology of China (USTC). He is currently a teacher in the Master of Finance program at USTC. By placing the lab in Hefei, they can have better access to outstanding students at USTC and also show some loyalty to the alma mater. "The high-caliber universities and the scientific genes in Hefei are not to be underestimated," commented industry insiders. It is reported that in July 2023, Meng Xi Investment upgraded its internship fund in Hefei, Anhui. Previously, they also had internship positions in quantitative strategy research, AI researchers, high-frequency development engineers (C++), and data development engineers. Quantitative private equity has always valued talent. Heiyi Asset stated that the company has systematically cultivated various talents and created three major recruitment programs: the "Fu Yao Plan" for interns, the "Yu Yi Plan" for campus recruitment of global graduates, and the "Kun Peng Plan" for mature top talents in social recruitment. Regarding talent retention, Heiyi Asset introduced that besides providing competitive salary and benefits, they also pay special attention to the happiness of employees, including providing basic benefits such as salary, bonuses, project incentives, and health insurance. In addition, they focus on the comprehensive development of employees by providing a quality working environment, continuous training courses, and the latest technology equipment. They also design a clear career development path for employees so they can effectively plan their career prospects. 400 billion giant Kuande WILL Lab synchronously recruits Similarly, Kuande Investment, a 400 billion giant, released a tweet on February 24 announcing the recruitment of talents for its intelligent learning laboratory, WILL Lab. The path of Kuande WILL Lab is similar to DeepSeek. The company was established with the strategic thinking about AI, and with the support of Kuande Investment, WILL will operate as an independent incubation, focusing on being an entrepreneurial laboratory for research in the super technology assistant field. According to the recruitment announcement, WILL Lab focuses on recruiting researchers and engineers with solid AI technical skills and research ideals. They hope to "jointly participate in this long-term intelligent research journey that requires continuous input." In terms of AI technology accumulation and development planning, Kuande Investment stated that they have made systematic investments in the field of quantitative research for many years, establishing complete AI infrastructure and data processing capabilities. WILL Lab will continue the excellent genes of Kuande Investment, starting from quantitative but not limited to financial scenarios and setting sail towards the vast ocean of artificial intelligence. In the golden age of AI development, a mature technical team can provide solid support for AI research and development, and a perfect talent cultivation mechanism can support innovative iterations. This is also the reason why Kuande Investment is recruiting talents. Jiukun, Mingcheng, and other leading companies have also made layouts. Although many quantitative private equity firms stated that their announcements about AI layouts may be seen as 'riding the wave,' the market still pays attention to these developments.Quantifying the information in the AI domain.Nine Kun Investment and Microsoft Asia Research Institute recently published an article stating that they have successfully reproduced DeepSeek-R1 for the first time, especially its achievements in the field of reinforcement learning, and have proposed innovative insights at the technical level. The academic article is titled "Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning," and was co-authored by Microsoft Asia Research Institute, Ubiquant, and other independent researchers. According to the paper, the team also first discovered issues such as the irrelevant improvement of reasoning performance with output length, significant reduction of reasoning ability with language mixing (such as mixing Chinese and English), and the actual increase in reasoning performance with reasoning tokens. It is reported that Nine Kun Investment, also a quantitative giant, established an AI lab early on and has strong technical and talent reserves in data, algorithms, and computing power. Since 2020, it has successively established AI labs, data labs, and droplet labs corresponding to research in the fields of data, algorithms, and transaction execution, respectively. In addition, the company also collaborated with the Guangdong-Hong Kong-Macao Greater Bay Area Digital Economy Research Institute in 2021 to establish the "Nine Kun-IDEA" joint laboratory, exploring new models for cooperation and development in the digital financial industry. In recent years, Nine Kun Investment has been actively conducting systematic and in-depth research in the field of cutting-edge AI technology, not only exploring general technologies and promoting their application in various scenarios, but also conducting diversified research in multiple sub-fields, striving to build a more comprehensive AI technology system. Minghong Investment also has corresponding reserves in AI technology. In 2020, Minghong Investment established a research center in North America to provide world-class technical support for A-share stock selection models. The speed of model algorithm iteration is based on high-performance computing power. From early simple CPU servers to large-scale high-performance computing clusters, Minghong has been actively building high-spec computing rooms, further improving supercomputing processing capabilities and levels. At present, Minghong Investment's own high-performance computing cluster has thousands of GPU cards, tens of thousands of CPU cores, stacked memory and disk storage of multiple Pb, and the AI computing power in financial data applications can reach 400P Flops, ranking on the world's TOP500 supercomputer list. What talents do quantitative giants favor? Blackwing Asset, a billion-dollar private equity firm, has been laying out in the field of artificial intelligence since 2017 and has formed an AI algorithm team, focusing on training and reserving talents in data analysis and machine learning. Even though they have not established a dedicated AI lab, they have achieved full AI integration in the process of quantitative investment. "In terms of hiring standards, we prefer AI talents who have a deep understanding of machine learning and deep learning technologies, and are filled with passion and curiosity for AI," said a representative from Blackwing Asset. Having rich research experience in well-known AI-related laboratories, research institutes, and companies at home and abroad, with a wealth of research achievements, publications in top international conferences or journals, or awards in algorithm competitions such as ACM/IOI/NOI/Top Coder/Kaggle will be more favorable. Pansong Asset stated that in recent years, the quantitative industry has been recruiting artificial intelligence talents, especially in deep learning, mainly for three reasons: First, the exponential growth in data dimensions and complexity. Traditional quantitative models are finding it difficult to efficiently extract valuable signals from unstructured data, and there is an urgent need for AI technology to perform integrated analysis and feature extraction of multimodal data. Second, the deepening of market competition demands higher adaptability to strategies. The advantages of deep learning in modeling nonlinear relationships and dynamic pattern recognition can help strategies quickly capture changes in market microstructure. Third, the competition of technological barriers has escalated to a strategic level. The quant industry is gradually building a closed-loop ecology of "AI+Quant," requiring teams to have interdisciplinary collaborative abilities, and composite talents have become scarce resources. Just like Blackwing Asset, several interviewed quantitative private equity firms told the reporter from Cailian News that the application of AI in strategies is a common phenomenon in the industry. Pansong Asset stated that AI technology is currently mainly used in three ways: fine-tuning data to better characterize the systemic investment logic, empowering the investment process, and establishing an efficiency committee within the company to use AI technology to enhance the efficiency of daily work operations. Interestingly, there are not many talents with overseas backgrounds in the DeepSeek team. Market professionals are worried that overseas hedge funds do attract top talents due to their global brand, mature training systems, and more attractive salary structures, creating a "suction effect" for top talents. In the talent war, what should Chinese local institutions do? Pansong Asset believes that local institutions still have strong advantages for three reasons: First, a deep understanding of the local market and agile response capability. The Chinese capital market has certain characteristics in trading mechanisms and investor structure. Local institutions can have a deeper understanding of the investment logic of the A-share market. Systematic investment can more deeply explore the economic logic of market phenomena, establish more precise mapping relationships through historical experience and data accumulation, and research needs to penetrate the "interpretability" of economic logic and the "statistical significance" of models simultaneously. Local institutions have a natural advantage in long-term data sample accumulation and sensitivity training. This dual verification mechanism of "logic + data" can significantly enhance the confidence of factor mining and the sustainability of strategies, while overseas institutions often require a longer adaptation period. Second, the efficiency of scenario-based technology innovation. Domestic teams are closer to the characteristics of the local market in terms of strategy iteration speed and model fault tolerance mechanism design, which is crucial for the practical value conversion of AI talents. Third, organizational culture compatibility and long-term incentive design. Compared to overseas institutions, local private equity firms can adopt a flat decision-making mechanism, deep coupling of technology and investment (such as the dual-track promotion channel of "researcher-engineer"), and equity incentives.Incentives such as medium to long-term binding methods enhance the attractiveness of top talent.This article is reproduced from Cailianshe, edited by GMTEight: Chen Wenfang.