PeopleofSuperintelligence

Researcher

Hoo-Chang Shin

Also known as Hoo Chang Shin

Hoo-Chang Shin is a Senior Research Scientist at NVIDIA whose research runs from deep learning for medical images to clinical language models and reward models for aligning language models. After doctoral studies at the Institute of Cancer Research from 2009 to 2013, they spent 2014 to 2016 as a postdoctoral fellow at the NIH Clinical Center, where they were first author of the 2016 paper on convolutional neural networks for computer-aided detection. At NVIDIA, they trained a generative adversarial network for synthetic brain tumor MRI in 2018, published the BioMegatron biomedical language model in 2020, and co-authored the 2022 GatorTron paper on a clinical language model with 8.9 billion parameters. In 2025, they co-authored the HelpSteer3 and HelpSteer3-Preference papers, the RLBFF reward-model method and the Nemotron Nano 2 report.

Now at
NVIDIA

Timeline

  1. 2009–2013

    Institute of Cancer Research PhD

    From 2009 to 2013, they are a PhD student at the Institute of Cancer Research.

  2. 2014–2016

    Postdoctoral fellow at NIH

    At the NIH Clinical Center from 2014 to 2016, they work as a postdoctoral fellow in the Imaging Biomarkers and Computer-Aided Diagnosis Laboratory, part of the Department of Radiology and Imaging Sciences.

  3. 2016

    CNNs for computer-aided detection

    In 2016, they are first author of Deep Convolutional Neural Networks for Computer-Aided Detection, written with Holger R. Roth, Le Lu, Ronald M. Summers and colleagues, which compares CNN architectures, dataset scale and transfer learning on lymph node detection and interstitial lung disease classification. Learning to Read Chest X-Rays follows on arXiv in March 2016, a model that pairs convolutional and recurrent networks to annotate chest X-rays with diseases and their contexts.

  4. 2018

    Synthetic brain tumor MRI with GANs

    At NVIDIA in 2018, they train a generative adversarial network on two public brain MRI data sets to generate synthetic abnormal MRI images with brain tumors, together with colleagues from the MGH & BWH Center for Clinical Data Science and the Mayo Clinic. The synthetic images improve tumor segmentation as data augmentation and serve as an anonymization tool that gives comparable segmentation results to real subject data.

  5. 2020

    BioMegatron biomedical language model

    In November 2020, they publish BioMegatron: Larger Biomedical Domain Language Model in the proceedings of EMNLP, with Yang Zhang, Evelina Bakhturina, Raul Puri, Mostofa Patwary, Mohammad Shoeybi and Raghav Mani. The paper studies sub-word vocabulary, model size, pre-training corpus and domain transfer, and reports gains over the previous state of the art on biomedical named entity recognition, relation extraction and question answering.

  6. 2022

    GatorTron clinical language model

    In December 2022, npj Digital Medicine publishes A large language model for electronic health records, co-authored by them at NVIDIA with University of Florida researchers. Its GatorTron models scale a clinical language model from 110 million to 8.9 billion parameters, trained from scratch on more than 90 billion words of text.

  7. 2025

    Reward models and HelpSteer3

    In 2025, they co-author HelpSteer3, whose feedback and edit models support inference-time scaling on open-ended tasks, and HelpSteer3-Preference, an open dataset of over 40,000 human-annotated preference samples used to train reward models. They also co-author RLBFF, a reward-model method built on binary flexible feedback, and the NVIDIA Nemotron Nano 2 report on a hybrid Mamba-Transformer reasoning model.

  8. As of October 2026

    Senior Research Scientist at NVIDIA

    They are a Senior Research Scientist at NVIDIA, after earlier work as a postdoctoral fellow at the National Institutes of Health. Their listed fields are artificial intelligence, machine learning, computer vision and natural language processing.

Career

  1. NowNVIDIASenior Research Scientist
  2. 2014–2016National Institutes of HealthPostdoctoral Fellow

Education

  1. 2013Institute of Cancer ResearchPhD

From the biography.

  1. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  2. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  3. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  4. Universities (31%) and startups and other companies (25%) employ the largest shares of 2020–2026 entrantsIncludes NVIDIA. Among 819 people whose first recorded role started in 2020–2026 and who have a known current organization, 60% work in commercial sectors (25% at startups and other companies, 23% at frontier labs and 13% at big tech, with 9 in two), and 31% work at universities, compared with 66% for earlier entrants.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Hoo-Chang Shin yet. Send us a free-licensed photo.

  1. Hoo-Chang Shin | alphaXiv alphaxiv.org. Checked 21 October 2026.
  2. Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning - PMC pmc.ncbi.nlm.nih.gov. Checked 21 October 2026.
  3. [1603.08486] Learning to Read Chest X-Rays: Recurrent Neural Cascade Model for Automated Image Annotation arxiv.org. Checked 21 October 2026.
  4. [1807.10225] Medical Image Synthesis for Data Augmentation and Anonymization using Generative Adversarial Networks (ar5iv) ar5iv.labs.arxiv.org. Checked 21 October 2026.
  5. [1807.10225] Medical Image Synthesis for Data Augmentation and Anonymization using Generative Adversarial Networks arxiv.org. Checked 21 October 2026.
  6. BioMegatron: Larger Biomedical Domain Language Model - ACL Anthology aclanthology.org. Checked 21 October 2026.
  7. A large language model for electronic health records - PMC pmc.ncbi.nlm.nih.gov. Checked 21 October 2026.
  8. [2503.04378] HelpSteer3: Human-Annotated Feedback and Edit Data to Empower Inference-Time Scaling in Open-Ended General-Domain Tasks arxiv.org. Checked 21 October 2026.
  9. [2505.11475] HelpSteer3-Preference: Open Human-Annotated Preference Data across Diverse Tasks and Languages arxiv.org. Checked 21 October 2026.
  10. [2509.21319] RLBFF: Binary Flexible Feedback to bridge between Human Feedback & Verifiable Rewards arxiv.org. Checked 21 October 2026.
  11. [2508.14444] NVIDIA Nemotron Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Model arxiv.org. Checked 21 October 2026.
  12. Hoo-Chang Shin - Google Scholar scholar.google.com. Checked 21 October 2026.