Why DeepSeek's Mass Recruitment Signals a New Phase in China's AI Race
On June 25, 2026, DeepSeek posted its largest-ever recruitment announcement: plans to double the company's workforce from 300 to 600+ employees. This isn't just hiring—it's a strategic pivot from "research lab focused on core technology" to "full-scale AI company competing globally." And it's happening at a moment when DeepSeek has just been valued at $50 billion (340 billion RMB), ranking 12th globally among unicorns.
Key Takeaways
- Double workforce: 300 → 600+ employees across 33+ position categories
- $50 billion valuation: Ranked #1 among 2026 new unicorns globally
- New Agent team: Building "DeepSeek Harness" to compete with Anthropic's Claude Code
- AGI hiring: Special "Cross-Domain AI Talent" role for candidates who want to "build AGI"
The Numbers: DeepSeek's Workforce Evolution
To understand why this recruitment matters, let's examine DeepSeek's unique position in the AI industry:
DeepSeek's approach has always been efficiency over scale. While OpenAI employs over 5,300 people and Anthropic about 5,000, DeepSeek achieved comparable results with just 300 researchers. This "talent density" philosophy—pioneered by founder Liang Wenfeng—proved that smaller, more focused teams could outperform massive organizations.
So why change course now?
"We are working hard to double the scale of all departments. We want new joiners to participate in the most important tasks immediately. Many of our colleagues have rapidly grown into top-tier industry talent, becoming core forces in driving AGI development."
— DeepSeek Recruitment Notice
What's Being Hired: 7 Major Categories
The recruitment announcement covers 7 major departments with 33+ specific position types. Here's what DeepSeek is looking for:
1. Full-Stack Development/Algorithm
- Backend engineers
- Algorithm engineers
- Infrastructure specialists
- Security engineers
2. Core AI System Research
- AI system engineers
- Hardware-software integration
- Performance optimization
3. Operations
- Site reliability
- DevOps
- Data operations
4. Product Design
- Product managers
- UX/UI designers
- User research
5. Model Data Strategy
- Data scientists
- Training specialists
- Evaluation experts
6. Deep Learning Research
- Research scientists
- Paper authors
- Novel architecture explorers
7. Specialized Domains
- Healthcare AI PM
- Legal AI PM
- Multilingual PM
The Agent Play: Building "DeepSeek Harness"
Among the most notable new positions is the "Agent Harness" team—explicitly positioned as DeepSeek's answer to Anthropic's Claude Code. This team is tasked with building code agents that can "actually get work done," matching Anthropic's strategy of building agents that automate real coding tasks.
Why Code Agents Matter
Code generation and autonomous coding agents represent one of the most commercially valuable AI applications. Anthropic's Claude Code has become the dominant tool in AI-powered software development, with over $2.5 billion in ARR from enterprise customers using it. DeepSeek's move into this space signals:
- Direct competition: DeepSeek believes its models can match or exceed Claude for coding tasks
- Market validation: The AI coding market is large enough to justify dedicated product teams
- Talent acquisition: Anthropic has been aggressively hiring for Claude Code; DeepSeek wants to match that investment
The "AGI Enthusiast" Role: Unusual Hiring Strategy
Perhaps the most unusual aspect of DeepSeek's recruitment is the "Cross-Domain AI Talent" position—a role specifically for candidates who "possess extraordinary abilities beyond ordinary people and want to participate in building AGI." Crucially, this position has no professional background requirements.
"We never look for geniuses. As long as you have your own shining spot, you are the person we are looking for."
— DeepSeek's hiring philosophyThis approach reflects DeepSeek's founder Liang Wenfeng's belief that breakthrough AI research often comes from unexpected sources. Rather than requiring specific degrees or previous company affiliations, DeepSeek looks for demonstrated ability and passion for the work itself.
The Global AI Talent War: DeepSeek Joins the Fight
DeepSeek's recruitment doesn't happen in isolation. The entire AI industry is experiencing unprecedented talent competition:
Recent High-Profile Moves
- Wu Yonghui, Google DeepMind VP of Research, departed for ByteDance to lead their Seed foundation research team
- Jonas Adler and Alexander Pritzel, two core Google Gemini researchers, plan to leave for Anthropic
- Apple reportedly doubled engineering salaries to retain AI talent facing recruitment from startups
Salary Benchmarks: AI Scientists Lead
According to data from professional networking platform Maimai, AI scientist/lead positions now command the highest salaries in China's tech industry:
| Position | Average Monthly Salary |
|---|---|
| AI Scientist/Lead | 132,796 RMB (~$18,300) |
| Algorithm Engineer | 74,441 RMB (~$10,250) |
| Average Tech Position | ~40,000 RMB (~$5,500) |
The talent gap is widening: AI scientist salaries are now 3.3x the average tech position, creating enormous competition for the best researchers.
What This Recruitment Reveals About DeepSeek's Strategy
Analysis of the recruitment reveals several strategic signals:
1. From Research Lab to Product Company
The hiring of product designers, user researchers, and specialized domain PMs (healthcare, legal, multilingual) indicates DeepSeek is building capabilities beyond pure research. This suggests monetization and productization are becoming priorities.
2. Geographic Distribution: Beijing and Hangzhou
Positions in both Beijing and Hangzhou give DeepSeek access to China's two major tech talent pools—Beijing's research institutions and Hangzhou's e-commerce and application development ecosystem.
3. Infrastructure Scaling
DeepSeek recently raised 50 billion RMB and reportedly operates 30,000 GPU racks in Wulanchabu, Inner Mongolia. The operations and SRE hiring suggests they're preparing for significantly larger scale deployments.
4. International Expansion Preparation
The multilingual product manager position hints at plans to better serve international markets—currently a weakness compared to OpenAI's global presence.
DeepSeek's Journey: From 300 to 600+
2023: Company Founded
Liang Wenfeng establishes DeepSeek with focus on AGI research
2024: DeepSeek V2 Release
MoE architecture proves Chinese labs can match Western capabilities
January 2025: DeepSeek R1
Reasoning model sparks global AI race; proves efficiency-first approach
June 2025: DSpark Framework
Open-sources inference acceleration, helps entire industry
June 2026: Mass Recruitment
Announces doubling of workforce; $50B valuation confirmed
Industry Implications: What This Means for China's AI
DeepSeek's transformation from elite research lab to full-scale company has broader implications:
Broader Industry Impact
Talent market effects: DeepSeek's competitive salaries and high-profile projects will attract talent from ByteDance, Alibaba, Tencent, and Baidu. This creates a new "premium tier" employer in China's AI landscape.
Research democratization: DeepSeek's history of open-sourcing models (V3, R1, DSpark) suggests their growth will continue benefiting the broader AI ecosystem, not just DeepSeek's proprietary products.
Commercialization pressure: With 600+ employees and $50B valuation, DeepSeek faces pressure to show revenue. This may accelerate their move from research leader to commercial competitor.
How DeepSeek Compares: Small but Mighty
Despite growing to 600, DeepSeek will remain remarkably small compared to Western AI giants:
| Company | Employees | Valuation | Per-Employee Value |
|---|---|---|---|
| Anthropic | ~5,000 | $66B | $13.2M |
| OpenAI | ~5,300 | $300B | $56.6M |
| DeepSeek (current) | ~300 | $50B | $166M |
| DeepSeek (post-hire) | ~600 | $50B | $83M |
Even after doubling in size, DeepSeek maintains extraordinary talent density—nearly 3x the per-employee valuation of Anthropic and 1.5x OpenAI. This validates Liang Wenfeng's philosophy that in AI, talent quality matters more than quantity.
Conclusion: The Next Phase of China's AI Story
DeepSeek's mass recruitment announcement isn't just about filling positions—it's a statement of intent. After proving that efficient research could compete with massive compute investments, DeepSeek is now building the organization to compete at scale.
The move signals several things:
- Commercialization is coming: DeepSeek is no longer just a research lab; it's building the teams needed to compete commercially
- Agent race is real: The Agent Harness team confirms that autonomous AI agents are the next battlefield
- China's AI maturity: Companies like DeepSeek represent a new level of sophistication in Chinese AI development
- Global competition intensifies: With DeepSeek joining the talent war, the competition between US and Chinese AI companies will only accelerate
For observers of China's technology sector, DeepSeek's evolution from 300-person research lab to 600-person AI company marks an important milestone. The company that proved "you don't need billions to compete in frontier AI" is now growing up—and that growth will reshape the global AI landscape.