Le Yu
Le Yu is an algorithm engineer at Alibaba Group, where he works on alignment for the Qwen series of models. Earlier, he studied at Beihang University, first for a bachelor's degree and then for a PhD that ended in 2024. From June 2023 to August 2024, he was a research intern at Tongyi Laboratory. He co-authored the paper Gated Attention for Large Language Models, which was named a NeurIPS 2025 Best Paper. In May 2025, the Qwen3 Technical Report, which he co-authored, was submitted to arXiv.
Timeline
-
2015–2019
A bachelor's degree at Beihang
Starting in September 2015, he studies for a bachelor's degree at Beihang University.
-
2019–2024
A PhD at Beihang
From September 2019 to June 2024, he studies for a PhD at Beihang University, supervised by Weifeng Lv.
-
June 2023–August 2024
A research internship at Tongyi
Between June 2023 and August 2024, he works as a research intern at Tongyi Laboratory.
-
May 2025
Gated attention for large language models
Posted to arXiv in May 2025, Gated Attention for Large Language Models lists him as a co-author.
-
May 2025
The Qwen3 technical report
Submitted to arXiv in May 2025, the Qwen3 Technical Report lists him as one of its authors.
-
2025
A NeurIPS 2025 Best Paper
The NeurIPS 2025 awards name Gated Attention for Large Language Models a Best Paper, with Le Yu among its authors.
-
Since 2024
Aligning the Qwen models
In September 2024, he joins Alibaba Group as an algorithm engineer, responsible for alignment of Qwen series models during post-training.
Sources and credits
Photo: We have no published photo of Le Yu yet. Send us a free-licensed photo.