Career opportunity distribution in the AI era follows a pattern: careers closest to the AI technology itself (ML engineers, LLM researchers) require strong technical backgrounds; the middle layer between AI technology and users/business (AI Product Managers, AI application architects) requires technical understanding + product/business capability; the application and dissemination layer of AI capability (AI training consultants, AI content creators) has the lowest technical requirements but needs vertical domain expertise. All three layers have abundant opportunities with varying entry barriers and salary ceilings.
## AI Product Manager (AI PM)
AI Product Manager is currently one of the hottest AI emerging positions: responsible for planning and executing products that translate AI technology capability into actual user value. Core capabilities: ① deep understanding of AI technology capabilities and limitations (able to communicate effectively with engineers; able to judge which features are actually feasible); ② user needs insight (AI PMs must understand users’ actual pain points and use scenarios better than technical teams); ③ AI product-specific design thinking (how to handle AI’s uncertain outputs, how to design user trust-building pathways with AI, how to A/B test AI products). Salary: domestic major tech AI PM approximately RMB 500,000-1,500,000/year; top AI companies (ByteDance, Alibaba, Baidu, Tencent AI Lab related) can reach RMB 2,000,000+.
## Prompt Engineer
Prompt Engineer was briefly considered “the next essential career” in 2022-2023, but as LLM capabilities improve (models tolerate ambiguous instructions better) and more people develop basic Prompt skills, pure “Prompt Engineer” as a standalone position is evolving: the more common form is “embedding Prompt engineering skills within other roles” — data scientists with Prompt engineering capability, UX researchers with Prompt engineering capability, rather than specialized Prompt Engineer recruitment. Exception: large enterprise AI Centers of Excellence (AI COE) and AI-native companies (like Anthropic, OpenAI) still have dedicated Prompt Engineer positions, but competition is intense.
## AI Application Entrepreneurship
The AI application layer (wrapper applications built on OpenAI/Anthropic APIs for vertical niches) is the most active current AI entrepreneurship direction: education (AI tutoring, AI language learning), legal (AI contract review, AI legal consultation), healthcare (AI diagnostic assistance tools, AI health management), content creation (AI writing, AI design). Competition characteristics: relatively low technical barriers (based on general APIs); differentiation comes from vertical data, user experience, and domain expertise; the first wave of successes (Jasper, Copy.ai, Notion AI) have already demonstrated commercial viability.
See [Human-AI Collaboration Skills](https://sunqi.org/human-ai-collaboration-en/) and [a16z AI Market Map](https://a16z.com/ai/).
