I am Haoyu Dong, currently in the final year of my Ph.D. in Electrical & Computer Engineering at Duke University, advised by Prof. Maciej A. Mazurowski. My research interests center on large language models, multimodal agents, and foundation models. I am particularly interested in building agentic systems that reason across images and text, use specialized tools, and adapt to complex, real-world medical tasks.

Across academic and industry research, I have developed multimodal medical agents that integrate language, vision, and external tools; worked on radiology report generation and text-guided image retrieval; and applied LLMs to harmonize labels across large, heterogeneous medical datasets. I have also built MRI foundation models, studied SAM and SAM 2 for medical image analysis, and improved Meta-Harness by developing adaptable harnesses for diverse use cases. I previously earned my M.S. in Computer Science at Duke, where I worked with Prof. Guillermo Sapiro, and my B.S. in Mathematics–Computer Science and Cognitive Science at UC San Diego, where I worked with Prof. Zhuowen Tu.

I am on the job market and actively seeking an industry position starting in Summer 2027!

Experience

Software Engineering Intern, Meta · May 2026 – Present
Improved the Meta-Harness algorithm by developing diverse, adaptable harnesses tailored to different use cases.

Research Intern, Siemens Healthineers · June 2026 – August 2026
Developed SNAIL, a 3D in-context learning framework for CT segmentation.

Recent News

• 🔥 [Jan. 2026] Our work (PDF) on Fréchet Radiomic Distance (FRD) is published in Medical Image Analysis.
• 🔥 [Jan. 2026] Our work (PDF) on applying SAM 2 to 2D and 3D medical images is published in IEEE Transactions on Biomedical Engineering.
• 🔥 [Dec. 2025] Our work (PDF) on breast MRI registration is published in IEEE Journal of Biomedical and Health Informatics.
• 🔥 [Oct. 2025] Our work (PDF) on AI-based breast density quantification in MRI is published in npj Breast Cancer.
• 🔥 [Oct. 2025] Our work (PDF) on accelerating volumetric annotation with SAM 2 is published in IEEE Transactions on Medical Imaging.
• 🔥 [July 2025] Our work (PDF) on an MRI foundation model is out.
• 🔥 [May 2025] Our work (PDF) on SAM fine-tuning strategies is published in MELBA.
• [April 2025] Our work (PDF) on universal bone segmentation is published in Medical Image Analysis.
• [Dec. 2024] Our work (PDF) on evaluating medical image segmentation algorithms is published at NeurIPS Datasets and Benchmarks.
• [Oct. 2024] Our work (PDF) on building a multimodal medical agent is published in Findings of EMNLP.
• [June 2024] Our work (PDF) on anatomically controllable medical image generation is published at MICCAI.
• [March 2024] Our work (PDF) on test-time adaptation is accepted as an oral presentation at a CVPR workshop.
• [March 2024] Our work (PDF) on breast MRI tissue and vessel segmentation is published in Scientific Reports.
• [March 2024] Our work (Paper) on confidence-guided radiology report generation is published in Neurocomputing.
• [Sep. 2023] Our work (PDF) on anomaly detection on high-resolution images is accepted by IEEE TMI.
• [Aug 2023] Our work (PDF) on SAM evaluation on medical domain is accepted by MedIA.
• [April 2023] Our work (PDF) on pluralistic image completion is accepted by MedIA.

This Site Already Has Visitors