The impact of DeepSeek on the investment fund industry is becoming increasingly evident, especially visible in the rising demand for talent across various institutions. Since the beginning of February, a number of public funds have kicked off their spring recruitment efforts, initiating a new round of talent acquisition. Notably, amidst the wave of local deployments brought about by DeepSeek, there has been a noticeable increase in job postings related to artificial intelligence.
This surge in recruitment reflects a wider trend—DeepSeek has created significant ripples in the fabric of the fund industry. Several public funds have already started to reap the benefits offered by this technological infrastructure, generally acknowledging its undeniable influence in propelling the funds sector forward.
As institutions integrate this technology more widely and embark on digital transformation, there is a simultaneous emphasis on candidates' abilities in AI development and real-world application. The journey toward deeply integrating DeepSeek into the fund business cycle cannot be instantaneous; it requires long-term exploration and learning within the industry.
Positions related to algorithms and quantitative investment are garnering significant attention. For instance, on February 17, Huaxin Fund announced three openings for technical roles, including that of an AI Algorithm Engineer. According to their official job description, this position involves researching and applying mainstream artificial intelligence technologies such as machine learning and deep learning. Responsibilities range from developing and implementing AI algorithms to providing technical support for various project teams.
The complexity of this role necessitates that candidates possess a robust knowledge of machine learning and deep learning. The job posting specifies that applicants should "proficiently understand common machine learning and deep learning algorithms and tools," with preference given to those who have experience in developing applications for large models, such as ChatGPT or DeepSeek, along with substantial practical experience in training, accuracy optimization, and deployment.
Coincidentally, on February 11, Huitianfu announced a recruitment for a Senior IT Manager with a clear focus on AI applications. Candidates must be adept in Python programming and familiar with major machine learning frameworks while possessing a deep understanding of Transformer architectures and large language model principles, along with at least three years of experience in training and deploying large language models.
From a functional perspective, the Senior IT Manager's responsibilities are twofold. Firstly, the individual must oversee the end-to-end development and optimization of large language models—including data processing, model evaluation, pre-training, fine-tuning, and deployment—enhancing the model's operational efficiency and inference speed. Secondly, the applicant will also be responsible for the practical application of the model in product scenarios, designing and implementing solutions based on large language models to improve core business efficiencies in investment research, risk control, and market services.
E Fund also rolled out a campus recruitment initiative on February 10 for an algorithm researcher position aimed at doctoral candidates. Requirements included an understanding of machine learning algorithm principles and the ability to conduct modeling studies with business scenarios, collaborating with engineers to develop AI financial projects, while keeping track of the latest developments in AI research and industry applications.
In addition, smaller public funds are actively recruiting talent tightly connected with machine learning, focusing their recent hiring efforts on algorithmic and quantitative research positions. For example, on February 14, Hongde Fund launched a new campus recruitment drive that included algorithm researcher positions with the expectation that candidates monitor the latest developments in deep learning, distill investment-worthy model concepts, reproduce and optimize related academic papers, and develop groundbreaking neural network models for continuous improvement.
Meanwhile, Guolian Fund has explicitly stated in its 2025 spring internship recruitment notice that they seek quantitative researchers familiar with statistics and machine learning, with job tasks that include conducting quantitative stock selection model studies based on machine learning methodologies.
However, it is important to note that not all public funds are rushing to bolster their technical talent pools. Some companies have already laid the groundwork in model development and application, while others are hesitant to invest further until they can ascertain the tangible benefits models will bring.
A representative from a mid-sized public fund mentioned to a reporter that the company currently has no new recruitment movements, likely due to their technical department having started AI-related research and layout much earlier, and their development team having previously allocated more personnel to support AI initiatives. This firm has been exploring the integration of AI into their business processes and had already taken steps for private model deployment before integrating DeepSeek.
Another public fund that has initiated its DeepSeek deployment planning shared their perspective, stating that as the industry is not yet in an expansion phase, they will follow the industry trend and consider hiring AI-related talent subsequently. Influences on their future operational decisions include the results of AI model applications and the cost-benefit analysis of their investments.
The intense pursuit of talent by fund companies is intricately linked to the transformative changes being ushered in by the implementation of DeepSeek. Since the Lunar New Year, many public funds have publicly expressed their progress in the deployment and application of DeepSeek, and this trend has continued into mid to late February.
On February 16, Caitong Fund and Western Leading Fund announced that they had completed local deployments of DeepSeek, and prior to this, over ten fund companies had announced their integration with the large model.
Some leading fund companies, while not explicitly stating their plans regarding DeepSeek, disclosed to reporters that they had already initiated local implementation of the platform. Unlike connecting via links, local private deployments not only ensure data security and response efficiency but also enhance the performance and stability of model services.
In fact, many public funds have previously utilized AI to empower their operations. Since the beginning of 2024, numerous medium to large public funds have integrated multiple mainstream AI models into their company systems and even explored self-developed models. Compared to the open-source models of the past, DeepSeek's significantly lower training and inference costs, coupled with its leading availability and efficiency, have catalyzed the adoption of AI large models in the fund industry.
Currently, some fund companies are already experiencing the optimization benefits brought by DeepSeek in their operations. For example, a representative from Pengyang Fund described how the marketing material compliance review system developed on the Coze intelligent platform achieved over 85% rule hit rate but exhibited room for improvement when processing complex clauses. Recently, by introducing the DeepSeek model to enhance the original plan, its exceptional comprehension abilities have notably improved the precision of rule identification, with the automated review accuracy rising to over 95%. The model will continue to be optimized for its adaptability to financial terminology.
E Fund also mentioned that since adopting DeepSeek, they have leveraged the platform's excellence in large model synthetic data and knowledge distillation, resulting in significant upgrades and optimizations for their self-developed financial model, EFundGPT, with a particular focus on enhancing the expert framework and critical thinking capabilities.
The merger with DeepSeek-R1 has enabled EFundGPT to function more effectively in scenarios requiring extensive logic, including intelligent investment research, material audits, customer service, intelligent investment advising, and personal assistant roles.
Similarly, Fuguo Fund pointed out that after exploratory validations by their technology team, their localized deployment model has reached usable stages across internal data processing, code assistance generation, text generation, enterprise-grade RAG, and research report interpretation.
It is crucial, however, to recognize that although the fund sector's exploration of DeepSeek has witnessed a surge akin to "explosive" growth, achieving in-depth integration with business chains and transitioning to company benefits is not an instantaneous process.
Currently, applications of DeepSeek in marketing, customer service, compliance, and operational management are being realized comparatively swiftly, showing improvements in efficiency and accuracy in institutional tasks related to content production.
In contrast, the practical utility of DeepSeek in investment research is expected to require relatively more time, primarily due to development challenges, employee adaptability, and the validation cycle. As noted by the earlier mentioned Pengyang Fund IT department representative, DeepSeek's application in research and investment remains a high barrier to entry. The model's uses in research include semantic analysis of research reports, event attribution analysis, remodelling quantitative strategy development, and realizing efficient factor digging and back-testing optimizations.
A stakeholder from a large public fund echoed that although their investment research staff consciously tries to utilize AI in their work as much as possible, much of it remains at a personal practice stage, with numerous researchers still feeling apprehensive about relying on DeepSeek for stock selection. Conversely, the quantitative team finds it easier to engage with DeepSeek, having begun attempts at model training, although their progress remains unclear.
While local deployment of DeepSeek may not be difficult, the pressing question remains as to how to leverage AI to provide solutions tailored to specific business scenarios. This should extend beyond simply implementing technology within a company's internal system, aiming instead to create differentiated outputs and services for clients.
This ambition aligns with the current trend observed among several funds during recent AI recruitment, where candidates are expected to demonstrate proficiency in both technological development and the practical application of their expertise.