DeepSeek V4: The Future of AI Unveiled by Tsinghua Talent
Gu Yuxian, a doctoral researcher at Tsinghua University and winner of the university's 2025 special scholarship, has officially joined DeepSeek, the Chinese AI startup now valued at up to $59 billion. His name appeared on the author list of the DeepSeek V4 technical paper, confirming his involvement with the company's flagship model.
A Prized Recruit
Gu is best known for his work on MiniLLM and efficient large language model training, with nearly 5,000 citations on Google Scholar. His research focus on model distillation — a technique for compressing the capabilities of large models into smaller, faster ones — aligns with DeepSeek's emphasis on cost efficiency. According to Pandaily, Gu ranks among Tsinghua's most promising young AI researchers.
DeepSeek's Hiring Spree
The recruitment comes amid a broader push by DeepSeek to at least double the size of every department. The company announced in late June that it was hiring for 33 positions across seven categories, including full-stack development, algorithms, and research, as well as functional roles in human resources, legal, and finance. The South China Morning Post reported that the firm is seeking both experienced hires and newcomers as it pursues artificial general intelligence.
The hiring drive follows DeepSeek's maiden funding round of approximately 50 billion yuan ($7.4 billion), with Tencent and CATL emerging as the largest outside investors, according to Reuters. The round could value the company between $52 billion and $59 billion.
V4 and What Comes Next
DeepSeek released a preview of its V4 model on April 24, offering two variants — V4-Pro and V4-Flash — both supporting a one-million-token context window. A full official release is expected in mid-July 2026. The model has already drawn attention from U.S. evaluators; the National Institute of Standards and Technology confirmed V4 as the most capable AI model from China that it has assessed to date.
