Qwen2.5-Coder Technical Report

Junyang Lin (Justin Lin) is a researcher and open-source maintainer known for the Qwen family of models. His public profiles list interests in LLMs, AI agents, multimodal learning, long-horizon reasoning, world models, and reinforcement learning; multiple March 2026 news reports said he stepped down from the Qwen tech lead role.

Senior algorithm expert at Alibaba Group working on large language models, multimodal large language models, and diffusion models.

Research scientist at Meta in New York City and research advisor at the UCLA NLP group; previously completed a PhD in computer science at UCLA.

PhD student at The Hong Kong University of Science and Technology (Guangzhou) whose research interests include large language models, vision-language models, AI agents, and multimodal retrieval.

Researcher on Alibaba's Qwen team focused on large language models and NLP, with public research profiles listing a Nankai University background.

Research scientist in Tongyi Lab whose official profile highlights post-training and multimodal large language models.

Research scientist in Tongyi Lab whose official profile highlights work on efficient reinforcement learning, generalization, inference-time scaling, and reasoning for large language models.

Qwen researcher and co-lead whose work focuses on pretraining and post-training, multimodal models, agent systems, and large-scale model infrastructure.

Canonical link

Junyang Lin

Shuai Bai

Zeyu Cui

Jinze Bai

Kai Dang

Xiaodong Deng

Wenbin Ge

Chang Zhou