Looking at talent management from the departure of Qwen (千问) technical lead: top talent is “not managed out”

Departure and industry backdrop

It has been reported that Lin Junyang (林俊旸), the technical lead behind the Qwen (千问) large language model, has left his role — a departure that rekindles debate over how China’s AI firms keep senior engineers. The move arrives amid a global scramble for AI talent since ChatGPT’s breakthrough in 2022, when startups and incumbents alike have raced to field competitive models. For Western readers: China’s AI ecosystem is expanding rapidly, but it operates under additional pressures — from U.S. export controls on advanced semiconductors to broader tech decoupling — making human capital arguably the most flexible strategic asset remaining.

Why top technologists leave — and what firms miss

Fudan University School of Management (复旦大学管理学院) associate professor Liu Shengming (刘圣明) told readers that departures are often structural rather than merely personal. He points to a multi‑factor mismatch: sheer supply shortage (demand for AI roles outstrips supply by several times), the fast pace of model and business‑model change, and four concrete frictions — pay and equity, management style, availability of supportive resources, and strategic or ideological misalignment. Reportedly, senior technologists sometimes leave because their technical ideals collide with short‑term commercial imperatives or rigid, KPI‑driven culture. Why stay where you cannot pursue ideas at scale?

Management lessons: cultivate an ecosystem, don’t “manage out”

Liu argues the remedy is ecological, not authoritarian. Top technical teams, he says, are “grown” not coerced: leaders must become service‑oriented resource coordinators, learn enough technical language to earn respect, adopt flexible evaluation systems (OKR over crude KPIs), provide dynamic equity and patent‑linked incentives, and protect psychological safety so risky research can thrive. In a geopolitically fraught era — where access to chips and overseas collaboration can be constrained — the ability to bind top talent through long‑term incentives and a permissive R&D culture is also a national competitiveness issue.

Can China’s tech firms shift from command‑and‑control routines to ecosystems that let genius breathe? The question is urgent. As Liu notes, losing a star is rarely just a personnel problem; it often signals deeper mismatches between what elite technologists need to do and what their employers expect them to deliver.