Chinese artificial intelligence start-up DeepSeek has launched a significant recruitment drive aimed at expanding its back-end engineering team to address increasing demands on its computing infrastructure. The Hangzhou-based company announced plans to hire approximately 150 senior back-end engineers as it seeks to upgrade and maintain its systems amid rapid growth in user activity and data processing requirements.
Cui Tianyi, who joined DeepSeek in March from a quantitative trading background to lead the Harness team, highlighted the intensifying operational pressures due to surging data volumes, an expanding server base, increased training workloads, and a growing active user count. In a social media post, Cui emphasized that scaling computing systems leads to increased complexity, necessitating a comprehensive overhaul of the company’s back-end architecture to support new technical developments.
This focus on infrastructure engineering reflects DeepSeek’s shift from a primarily research-oriented entity to a large-scale commercial platform. The company is enhancing its systems to meet enterprise-grade reliability standards and facilitate high-volume service delivery. Earlier this year, DeepSeek revealed plans to expand all departments significantly, opening 33 new positions across research, engineering, and product management.
Financially, the company appears to be gaining momentum. It recently increased fees for its application programming interface (API) amid intense domestic competition and is reportedly close to completing a funding round that could value the firm at around 500 billion yuan (HK$584 billion) before the investment. This funding round precedes a potential public listing on Shanghai’s Star Market.
DeepSeek faces complex operational challenges tied to the rising use of autonomous AI agents, which require repeated model invocations and dedicated computing environments. These factors drive substantial capital expenditures on hardware, networking, and software, even as the firm strives to maintain competitive API pricing against both domestic and international rivals.
The 150 new hires will be distributed across two principal technical domains. One group will focus on server-side development encompassing large-model research platforms, agent frameworks, internal R&D infrastructure, public API maintenance, online services, and data engineering. The other will work on elastic computing infrastructure, splitting efforts between maintaining the agent platform and optimizing lower-level systems to enable flexible computation workloads.
The recruitment announcement coincided with DeepSeek initiating beta tests for its new DeepSeek V4.1 Flash model. According to a company notice shared on WeChat, this latest iteration features a redesigned architecture with native multimodal support. The firm described the V4.1 Flash as stronger, faster, and more cost-efficient and encouraged users to trial the model and provide feedback.