PeopleofSuperintelligence

Researcher

Byung-Kwan Lee

Also known as ByungKwan Lee

Byung-Kwan Lee is a research scientist at NVIDIA, where he builds efficient vision-language models. He studied mathematics and electronic engineering at Hanyang University, then earned master's and doctoral degrees at KAIST. During an NVIDIA Research internship, he worked on VLsI, a layer-wise distillation method for smaller vision-language models, and coauthored RIL, which combines reinforcement learning with adversarial imitation. He joined NVIDIA as a Research Scientist in October 2025, and his 2026 work includes recursive reasoning for language and vision-language models.

Now at
NVIDIA
Works
16
Citations
148
h-index
5
Badges

Publication numbers from OpenAlex, as of 19 October 2026.

Timeline

  1. 2014–2018

    A mathematics and engineering degree

    He earns a B.S. in mathematics and electronic engineering at Hanyang University in 2018. His undergraduate thesis uses a convolutional neural network to rank songs.

  2. 2018–2025

    Graduate study at KAIST

    He completes an M.S. in the Cho Chun Sik Graduate School of Green Transportation at KAIST in 2020 and earns a Ph.D. in electrical engineering in 2025. His dissertation studies high-performing, efficient vision-language models.

  3. October 2024–October 2025

    An NVIDIA Research internship

    He interns at NVIDIA Research from October 2024 to October 2025 while working on VLsI, a layer-wise distillation method that helps smaller vision-language models align with larger models' reasoning. The paper appears at CVPR 2025.

  4. 2025

    Unified reinforcement and imitation

    With four coauthors, he presents Unified Reinforcement and Imitation Learning (RIL) in the NeurIPS 2025 main conference. The method combines reinforcement learning with adversarial imitation to train lightweight vision-language models.

  5. 2025

    A Samsung HumanTech award

    In 2025, he receives recognition in the 31st Samsung HumanTech Paper Awards in Computer Science & Engineering.

  6. Since October 2025

    Research Scientist at NVIDIA

    He joins NVIDIA as a Research Scientist in October 2025 and focuses on high-performing, efficient vision-language models. In 2026, he coauthors Recursive Think-Answer Process for LLMs and VLMs, which uses iterative reasoning cycles and a confidence generator to improve answers.

Career

  1. Since 2024NVIDIAIntern, Research Scientist

Education

  1. 2025KAISTM.S.; Ph.D.
  2. 2018Hanyang UniversityB.S.

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Dimensions of Interactivity: Differential Effects of Social and Psychological Factors2006 · Journal of Computer-Mediated Communication · 83 citations
  2. How Attitude Toward the Web Site Influences Consumer Brand Choice and Confidence While Shopping Online2006 · Journal of Computer-Mediated Communication · 29 citations
  3. The Lightweight Autonomous Vehicle Self-Diagnosis (LAVS) Using Machine Learning Based on Sensors and Multi-Protocol IoT Gateway2019 · Sensors · 13 citations
  4. An Audification and Visualization System (AVS) of an Autonomous Vehicle for Blind and Deaf People Based on Deep Learning2019 · Sensors · 7 citations
  5. CoLLaVO: Crayon Large Language and Vision mOdel2024 · 5 citations

From OpenAlex, as of 19 October 2026.

  1. 14% of AI career moves in the 2020s went from a university to a companyIncludes Byung-Kwan Lee. Our estimates of how university-to-company career transitions changed by decade, including roles added alongside an earlier job.
  2. Google alumni founded or led 81 other organizations, the most among the labs we compareIncludes NVIDIA. We trace 194 organizations founded or led by AI lab alumni, and 202 career moves between the labs.
  3. 44% of AI career moves go from one university to anotherIncludes NVIDIA. We estimate how 13,335 AI career moves link universities, companies, labs and government research centers, with net flows and the most common paths.
  4. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. Our estimates of combined and median h-index at AI organizations, using current affiliations and OpenAlex records.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Byung-Kwan Lee yet. Send us a free-licensed photo.

  1. BK Lee byungkwanlee.github.io. Checked 18 October 2026.
  2. IVYLab & IVLLab - Alumni ivylab.kaist.ac.kr. Checked 18 October 2026.
  3. CVPR 2025 Open Access Repository openaccess.thecvf.com. Checked 18 October 2026.
  4. Unified Reinforcement and Imitation Learning for Vision-Language Models papers.nips.cc. Checked 18 October 2026.
  5. Byung-Kwan Lee | Research research.nvidia.com. Checked 18 October 2026.
  6. Recursive Think-Answer Process for LLMs and VLMs | NVIDIA Research Taiwan research.nvidia.com. Checked 18 October 2026.