Scholar
Jaeyoung Do
Google Scholar ID: eIMVw2MAAAAJ
Department of Electrical and Computer Engineering, Seoul National University
Citations & Impact
Citations
1,298
H-index
15
i10-index
19
Publications
20
Co-authors
0
Contact
Publications
Resume (English only)
Academic Achievements
- Published multiple papers including 'Don’t Let It Fade: Preserving Edits in Diffusion Language Models via Token Timestep Allocation', 'Exploring and Leveraging Class Vectors for Classifier Editing', 'MMPB: It’s Time for Multi-Modal Personalization', 'Turbocharging Vector Databases using Modern SSDs', 'SECOND: Mitigating Perceptual Hallucination in Vision-Language Models via Selective and Contrastive Decoding', 'MathSpeech: Leveraging Small LMs for Accurate Conversion in Mathematical Speech-to-Formula', 'MathReader: Text-to-Speech for Mathematical Documents'.
Research Experience
- Worked as a Senior Applied Scientist at Amazon Alexa AI from 2021 to 2023, and as a Senior Researcher at Microsoft Research (MSR) from 2016 to 2021. Held positions as a Senior Scientist at Microsoft Jim Gray Systems Lab (GSL) from 2014 to 2016, and as a Research Engineer at Microsoft Database Lab from 2012 to 2014.
Education
- Received his Ph.D. in Computer Science from the University of Wisconsin-Madison in 2012 under the supervision of Professor Jigensh M. Patel. Obtained his Master's degree from the same university in 2009 and earned his Bachelor's degree from Korea Advanced Institute of Science and Technology (KAIST).
Background
- Currently serves as an assistant professor in the Department of Electrical and Computer Engineering at Seoul National University (SNU). Research interests include generative AI based on Large Language Models (LLMs), Natural Language/Vision processing through multi-modal AI, Agent/Embodied AI focusing on physical interactions and robotic action planning, AI-powered analytics and query processing/optimization for big data systems, algorithm-system co-design for ML/AI applications, high-performance large-scale AI using next-generation memory and cutting-edge hardware technologies, and large-scale deep-learning models and infrastructure for healthcare and manufacturing industry.
Miscellany
- Looking for motivated undergraduate and graduate students interested in generative AI, multi-modal AI, agent/embodied AI, AI-powered analytics and query processing/optimization, algorithm-system co-design, high-performance large-scale AI, and large-scale deep-learning models for healthcare and manufacturing.
Co-authors