AI Researcher & Engineer at BAAI
I build dependable foundation models for specialized domains through post-training, retrieval agents, and multimodal reasoning.
ICML 2026 research · Open models, data, and retrieval tools.
Work: xfshi@baai.ac.cn
Three projects, with results and reusable resources.
Introduces the MechVQA mechanical-drawing benchmark and the MechVL domain-specialized multimodal model, with public code, benchmark data, and model resources.
Result: MechVL-4B-RL scores 84.85 on MechVQA Total, 7.57 points above Gemini-3-Pro-Preview (77.28) in the paper’s evaluation.
Internalizes document knowledge for retrieval-free question answering while recovering general model capabilities.
Result: Across eight dataset-model settings, domain QA accuracy improves by an average 3.6 percentage points over Vanilla SFT.
Introduces a multi-agent framework for scholarly retrieval and the SPARBench evaluation dataset with expert-reviewed relevance labels.
Result: F1 reaches 0.3843 on AutoScholar and 0.3015 on SPARBench: relative gains of 56.9% and 23.0% over the best baselines in Table 1.
我在智源研究行业大模型，围绕领域知识学习、检索智能体与多模态推理开展研究与工程合作。
