Senior Algorithm Engineer | Ant Group

Senior Algorithm Engineer at Ant Group. Currently working on LLM post-training (reasoning SFT data, DeepResearch SFT data) and small-scale model pre-training (~10B). Previously worked on knowledge graph reasoning (designed the SPG-Reasoner in OpenSPG), search & recommendation, and online learning systems. Earlier career includes backend development at Tencent (gaming) and SEA Ltd. (e-commerce).

M.Eng. in Computer Science
University of Electronic Science and Technology of China (UESTC)
Graduated 2011

• Ant Group — Senior Algorithm Engineer (2017 – Present)
  • LLM post-training & pre-training, knowledge graph reasoning (SPG-Reasoner), search & recommendation
• SEA Ltd. — Senior Backend Engineer (2015 – 2017)
• Tencent — Game Backend Developer (2011 – 2015)

• LLM Post-training — SFT data curation, reasoning data optimization, alignment
• Small Language Model Pre-training — Efficient pre-training at the ~10B scale
• Knowledge Graph Reasoning — Symbolic reasoning, neural-symbolic integration, KGDSL
• LLM Evaluation — Benchmark-driven model improvement

• OpenSPG — Designed the SPG-Reasoner module for logical rule reasoning with KGDSL
• Ling (百灵) Model Post-training — Post-training data optimization for Ant Group's trillion-parameter LLM family
• Small-scale LLM Pre-training — Pre-training ~10B-parameter language models from scratch

• GitHub: github.com/youdonghai
• ORCID: 0009-0009-0344-4735
• OpenSPG: github.com/OpenSPG/openspg
• Ling 2.6 Technical Report: arXiv:2606.15079

Last updated: July 2026