Weicai Long

About Me

I am a third-year (Final-year) Ph.D. candidate in Data Science and Analytics at HKUST(GZ), advised by Prof. Yanlin Zhang and Prof. Jiaheng Wei. Before joining HKUST(GZ), I received my M.Eng. in Electronic and Communication Engineering from Tsinghua University in 2023, advised by Prof. Xingjun Wang, and my B.Eng. in Communication Engineering from the University of Science and Technology Beijing in 2020.

My research focuses on BioMedical AI, spanning genome language models (gLMs) and medical large language models (MedLLM). I study domain-specific pretraining and multimodal representation learning for biomedical data, together with post-training, reasoning, evaluation, and agentic systems for healthcare. My work also draws on experience in wearable devices, with the goal of developing reliable AI systems for both patients and healthcare professionals.

I am currently on the job market and open to opportunities in both academia and industry. Please feel free to contact me at longwc819@gmail.com.

Research Interests

  • BioMedical foundation models. Pretraining and representation learning for biological sequences and heterogeneous biomedical data, including genome language models, and multimodal modeling of physiological signals, laboratory tests, and clinical information.
  • Medical LLM post-training. Domain-specific data construction, SFT, RL, and evaluation for improving medical reasoning, factuality, safety, and reliability.
  • Agents for healthcare systems. Tool-augmented and workflow-oriented agents for patient- and clinician-facing healthcare applications, with an emphasis on information gathering, multi-step task execution, long-term interaction, and effective human–AI collaboration.

To Prospective Collaborators

I am always open to collaboration with researchers and students interested in LLMs for biomedicine. I particularly welcome discussions on genome language models (gLMs), Medical LLMs, LLM post-training and AI agents for scientific workflows.

If you are interested in working with me, please feel free to email me at with a brief introduction to your background and research interests.

Talks

  • I give an oral presentation at ISMB/ECCB 2025. Video Link.

News

  • BioMaster is published in Patterns.
  • GenomeQA and PhageBench accepted to ACL 2026.
  • Porcine MutBERT is published in Briefings in Bioinformatics.
  • NyxBind is published in Briefings in Bioinformatics.
  • MutBERT accepted to ISMB 2025.
  • Started my Ph.D. in Data Science and Analytics at HKUST(GZ).
  • Started working full-time at HONOR as a Sport Health Algorithm Engineer.
  • BPNet is published in Biomedical Signal Processing and Control.
  • BP-LKA accepted to ICDIP 2023.
  • Started an internship at HONOR, working on cuffless blood-pressure estimation.

Selected Publications

[Full List]

† Equal contribution.

  1. GenomeHarness: Harnessing AI Agents for Reliable Adaptation of Genome Language Models [link]

    Weicai Long, Yusen Hou, Houcheng Su, Junning Feng, Yanlin Zhang

    Bioinformatics, 2026.[CCF-A · JCR-Q1 · IF: 5.5 · Under review]

  2. GenomeQA: Benchmarking General Large Language Models for Genome Sequence Understanding [link]

    Weicai Long†, Yusen Hou†, Junning Feng, Houcheng Su, Shuo Yang, Donglin Xie, Yanlin Zhang

    ACL 2026 · Main Conference, 2026.[CCF-A]

  3. PhageBench: Can LLMs Understand Raw Bacteriophage Genomes? [link]

    Yusen Hou†, Weicai Long†, Haitao Hu, Houcheng Su, Junning Feng, Yanlin Zhang

    ACL 2026 · Findings, 2026.

  4. BioMaster: Multi-agent system for automated bioinformatics analysis workflow [link]

    Houcheng Su, Junning Feng, Yawen Lu, Yucheng Xu, Jinming Yang, Haojie Lu, Jixin Yang, Xu Yang, Sirui Xie, Weicai Long, Chengrui Wang, Yusen Hou, Tingyu Zhu, Yanlin Zhang

    Patterns, 2026.[JCR-Q1 · IF: 10.8]

  5. Porcine MutBERT: A Family of Lightweight Genomic Foundation Models for Functional Element Prediction in Pigs [link]

    Weicai Long†, Rong Zhou†, Wenkang Wei, Xiaoai Zhang, Shanshan Wu, Kui Li, Yanlin Zhang, Zishuai Wang

    Briefings in Bioinformatics, 2026.[CCF-B · JCR-Q1 · IF: 7.3]

  6. ContextTAD: Context-aware boundary learning for TAD calling from Hi-C contact maps [link]

    Weicai Long†, Yusen Hou†, Yanlin Zhang

    Nucleic Acids Research, 2026.[JCR-Q1 · IF: 15.0 · Under review]

  7. PopGenAgent: An Agent-Mediated Workflow for Population Genetics Analyses [link]

    Houcheng Su, Weicai Long, Junning Feng, Yusen Hou, Yanlin Zhang

    bioRxiv, 2026.

  8. NyxBind: enhancing deep neural representations for transcription factor binding site prediction via contrastive learning [link]

    Xu Yang, Qingfa Xiao, Yucheng Xu, Jixin Yang, Yusen Hou, Weicai Long, Miaojun Huang, Yanlin Zhang

    Briefings in Bioinformatics, 2026.[CCF-B · JCR-Q1 · IF: 7.3]

  9. MutBERT: Probabilistic Genome Representation Improves Genomics Foundation Models [link]

    Weicai Long, Houcheng Su, Jiaqi Xiong, Yanlin Zhang

    ISMB 2025 (Published in Bioinformatics), 2025.[CCF-A · JCR-Q1 · IF: 5.5 · ISMB: Top Conference]

  10. Large Language Models for Network Intrusion Detection Systems: Foundations, Implementations, and Future Directions [link]

    Shuo Yang, Xinran Zheng, Xinchen Zhang, Jinfeng Xu, Jinze Li, Donglin Xie, Weicai Long, Edith C. H. Ngai

    arXiv, 2025.

  11. BPNet: A Multi-Modal Fusion Neural Network for Blood Pressure Estimation Using ECG and PPG [link]

    Weicai Long, Xingjun Wang

    Biomedical Signal Processing and Control, 2023.[JCR-Q2 · IF: 5.7]

  12. Blood Pressure Estimation Neural Network Using Large Kernel Convolutional Attention [link]

    Weicai Long, Jingyuan Li, Xingjun Wang

    ICDIP 2023, 2023.

Experience

Sport & Health Algorithm Researcher, HONOR · Chipset and Algorithm Solutions Department

  • Led the development of a blood-pressure algorithm data pipeline spanning MIMIC-II, MIMIC-III, and HONOR's real-world wearable data from more than 10,000 participants, covering ECG/PPG denoising, peak detection, label construction, artifact removal, signal-quality assessment, and subject-independent data splits.
  • Built reusable algorithm modules for wearable physiological signals, including filtering and preprocessing, activity-state detection, signal-quality gating, temporal feature extraction, and multimodal ECG–PPG fusion.
  • Advanced continuous cuffless blood-pressure estimation through individual calibration and cross-population generalization, systematically evaluating PTT/PAT, multi-wavelength PPG, personalized regression, domain adaptation, and distribution-shift correction.
  • Contributed to the 24-hour continuous non-invasive blood-pressure monitoring feature shipped on HONOR Watch 5 Pro and developed a related patented method.

Sport & Health Algorithm Intern, HONOR · Chipset and Algorithm Solutions Department

  • Built a high-quality cuffless blood-pressure modeling dataset covering approximately 2,500 participants.
  • Developed and validated BPNet and BP-LKA, exploring multimodal ECG/PPG representation learning, multi-task blood-pressure prediction, large-kernel attention, and the integration of personal attributes to establish technical directions for subsequent product development.

Teaching & Service

  • Teaching Assistant, Design and Analysis of Algorithms, HKUST(GZ), 2025 and 2026.
  • Teaching Assistant, Pattern Recognition, Tsinghua University, 2021.

Languages

  • Chinese — Native speaker.
  • English — Fluent (professional working proficiency).

Patent

  • A Blood Pressure Prediction Apparatus and Device (granted), Xingjun Wang and Weicai Long, 2025, CN115836847B.
  • A Blood Pressure Measurement Method and Related Devices (granted), Weicai Long, 2025, CN119235280B.

Awards

  • Second Prize, Tsinghua Tongfang Cup Sensor Competition, 2017.
  • USTB People’s Scholarships, 2016–2018.