Particle.news
Download on the App Store

Tencent Taps Tsinghua’s Pang Tianyu to Lead Multimodal Reinforcement Learning for Hunyuan

The hire signals a renewed push to rebuild Hunyuan’s research engine under recently installed leadership.

Overview

  • Pang Tianyu said he will join Tencent on February 4, 2026 as chief research scientist for the Hunyuan model team and head of multimodal reinforcement learning.
  • Tencent stated that Pang will lead reinforcement-learning research for multimodal models with an initial emphasis on generative-model techniques.
  • On January 28, Tencent open-sourced Hunyuan Image 3.0 for image-to-image tasks, integrated it into the Yuanbao AI assistant, and reported first-tier standing on LMArena’s image-editing benchmark.
  • Pang holds a Tsinghua University PhD under Prof. Zhu Jun, has 70-plus first‑author papers at top venues with more than 15,000 citations, and previously served as a senior research scientist at Sea AI Lab in Singapore.
  • In December 2025, Tencent named former OpenAI researcher Yao Shunyu as chief AI scientist, and CEO Ma Huateng later described the Hunyuan effort as undergoing a “deep reconstruction” driven by talent recruitment.