Summary of Key Points
FDE (Field Deployment Engineers) have become a popular AI role that has gained momentum from Silicon Valley to China, often referred to as the "last mile" in the implementation of AI solutions. However, the approaches in China and the US differ significantly: In Silicon Valley, FDEs are highly skilled professionals who understand technology, business, and consulting, earning annual salaries in the millions; in China, they more closely resemble external contractors, responsible for fixing issues that arise after sales agreements are made, facing challenges such as internal conflicts, misaligned requirements, and short-sighted management. The author illustrates these realities through three case studies involving state-owned enterprises, private education institutions, and home improvement companies.
1. FDEs in China and the US: One as "High-End Consulting," the Other as "External Contractors"
In Silicon Valley, FDEs are viewed as crucial players in AI implementation—they must possess technical expertise (such as understanding AI models), business acumen (knowing where AI can add value to companies), and consulting skills (helping with transformation planning). Their salaries can reach 1.5 to 2 million RMB per year, with even OpenAI getting directly involved in FDE-related projects.
In China, however, the role has been distorted. Most FDEs work for cloud service providers (like Volcano or Tencent Cloud) or as agents, essentially handling post-sales support. Salespeople make unrealistic promises to customers (e.g., delivering 10 intelligent systems within six months), and FDEs are left to fix any problems that arise. The author notes that their actual income is low, often not even covering travel expenses, making it a far from glamorous job compared to the Silicon Valley counterpart.
2. FDEs in State-Owned Enterprises: Managing People Is More Important Than Technology
The author worked on a large-scale project for a state-owned enterprise. After the management signed the contract, they handed it over to middle-level managers who, afraid of making mistakes, asked the FDEs to figure out the requirements on their own. As a result, seven intelligent systems were created but could not be used effectively due to lack of clear criteria and business value. Worse still, the original project leader set up his own AI department, and the IT manager only wanted to complete the project quickly, relying on FDEs to create reports to convince management (e.g., by producing a demo with some fabricated data).
The core issue in state-owned enterprises is the complex power dynamics: no one wants to take responsibility. Management makes decisions without caring for details, middle-level managers push projects forward, and FDEs are caught in the middle, unable to connect with either management or the business teams. The author concludes that reporting skills are more important than technical expertise, as management focuses only on superficial results.
3. FDEs in Private Enterprises: Managers Focus on Reducing Costs or Increasing Revenue
Private enterprise managers have straightforward goals: they want to maximize profits or cut expenses. For example:
- Case 1 (Private Education Group): A professional team spent three months developing a comprehensive AI system, but the principals found it too complicated and decided to use a simplified version created by a 60-year-old principal in just three days. The simplified version (which only generated templates that needed manual filling) was more widely accepted.
- Case 2 (Home Improvement Company): The manager asked the FDEs to cut 3,000 employees, but AI could only replace some routine tasks (e.g., contract review), not entire roles. The project ended in loss, and the manager was still dissatisfied.
The main problem is that managers are eager for quick results and use AI as a tool to reduce costs, ignoring the fact that AI cannot yet replace entire jobs and that many companies lack the digital infrastructure necessary for effective implementation.
4. The Future of FDEs: Short-Term Value, Long-Term Transition
The author believes FDEs will still be valuable in the short term for helping companies overcome initial barriers to AI adoption but will eventually become obsolete:
- Large Companies Will Develop Their Own Capabilities: State-owned enterprises and large firms will establish internal AI teams, reducing the need for external FDEs.
- Problems with the Agency Model: The multiple intermediaries (salespeople → agents → FDEs) often lead to project failures due to unrealistic promises from sales.
- AI Tools Will Simplify: Future AI platforms will be more user-friendly, allowing companies to deploy them without FDE assistance.
In conclusion, the author argues that no matter how advanced AI becomes, it cannot overcome the inherent complexities of human interactions within organizations. Where there are people, there will always be issues such as blame-shifting and power struggles, which AI cannot resolve.
5. Tips for Those Considering a Career as an FDE
The author offers four key recommendations for those interested in this role:
- In State-Owned Enterprises: Focus on Building Relationships: Build good relationships with key decision-makers.
- Direct Communication with Decision-Makers: Try to work directly with the person who signs the contracts to avoid unnecessary red tape.
- Choose Companies with a Digital Foundation: AI cannot be implemented without proper data and infrastructure.
- Be Patient with Managers: Avoid clients who expect immediate results; AI implementation takes time.
In summary, being an FDE in China is not a glamorous role but often requires dealing with challenges. However, with the right approach, it can still be profitable during this transitional period.