⚙️ Engineering Experience

💼 Internships

Internship
Qwen Team

Research Intern @ Qwen Team, Alibaba Group / DAMO Academy
2024.08 – Present   |   China

  • Working on code LLM pre-training, post-training, data synthesis, evaluation, and engineering infrastructure for the Qwen-Coder series.
  • Contributed across the full training pipeline of Qwen2.5, Qwen3, Qwen3.5, and Qwen3-Coder-Next code models, covering data construction, benchmark design, model evaluation, and rollout workflows.
  • Built and supported multiple internal systems, including code execution sandboxes, large-scale data generation pipelines, model serving scripts, SFT data visualization tools, and agent-based data construction workflows.
  • Participated deeply in projects on code completion, frontend code generation, chart code generation, software engineering data synthesis, and agentic coding model training.

📄 Technical Reports

Tech Report
Qwen2.5-Coder

Qwen2.5-Coder Technical Report
Qwen Team, with Jiajun Zhang as a core contributor

  • Summary: This report introduces the Qwen2.5-Coder family and details its model scaling strategy, code-centric pretraining data, benchmark evaluation, and performance across generation, completion, reasoning, and repair tasks.

arXiv

📖 BibTeX
@article{hui2024qwen25coder,
  title   = {Qwen2.5-Coder Technical Report},
  author  = {Binyuan Hui and Jian Yang and Zeyu Cui and Jiaxi Yang and Dayiheng Liu and Lei Zhang and Tianyu Liu and Jiajun Zhang and Bowen Yu and Keming Lu and Kai Dang and Yang Fan and Yichang Zhang and An Yang and Rui Men and Fei Huang and Bo Zheng and Yibo Miao and Shanghaoran Quan and Yunlong Feng and Xingzhang Ren and Xuancheng Ren and Jingren Zhou and Junyang Lin},
  journal = {arXiv preprint arXiv:2409.12186},
  year    = {2024}
}
Tech Report
Qwen3-Coder-Next

Qwen3-Coder-Next Technical Report \ Qwen Team, with Jiajun Zhang (Speakn0w) as a major contributor

  • Summary: This report presents Qwen3-Coder-Next, an open-weight coding-agent model with 80B total parameters and 3B active parameters during inference, trained with large-scale verifiable coding tasks, executable environments, mid-training, and reinforcement learning.

arXiv Model Blog

📖 BibTeX
@article{cao2026qwen3codernext,
  title   = {Qwen3-Coder-Next Technical Report},
  author  = {Ruisheng Cao and Mouxiang Chen and Jiawei Chen and Zeyu Cui and Yunlong Feng and Binyuan Hui and Yuheng Jing and Kaixin Li and Mingze Li and Junyang Lin and Zeyao Ma and Kashun Shum and Xuwu Wang and Jinxi Wei and Jiaxi Yang and Jiajun Zhang and Lei Zhang and Zongmeng Zhang and Wenting Zhao and Fan Zhou},
  journal = {arXiv preprint arXiv:2603.00729},
  year    = {2026}
}

🌐 Open-Source Contributions

  • Qwen3-Coder  
    Major contributor under the handle Speakn0w to the official Qwen3-Coder repository, especially in benchmark evaluation code, training-related workflows, release support, and engineering tooling.

  • Qwen  
    Contributed to the broader Qwen open-source ecosystem through code-model evaluation, infrastructure support, and model-related tooling.

  • FluxLLM
    Core contributor to an efficient asynchronous and parallel API-calling library for large-scale model inference and request scheduling.