Scholar
Haebin Seong
Google Scholar ID: NSX1w-EAAAAJ
KAIST
Citations & Impact
Citations
15
H-index
2
i10-index
0
Publications
5
Co-authors
6
list available
Publications
Resume (English only)
Academic Achievements
- - [arXiv] D2E: Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI
- - [ICLR 2025] HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models
- - [ACL 2025] SafeRoute: Adaptive Model Selection for Efficient and Accurate Safety Guardrails in Large Language Models
- - [NeurIPS 2025] FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA
Research Experience
- - Maum.ai, 2025.03 ~ Current, Seoul, Korea, Senior AI Research Scientist, responsible for pre-training and post-training of foundational visual-language-action (VLA) models from scratch for autonomous driving and robotics; developing end-to-end generalizable & instructable robotic agents integrated for real-world agricultural vehicles; creating open-source projects for Open-World-Agents framework.
- - Theori, 2023.11 ~ 2025.03, Seoul, Korea, AI Engineer, providing consulting for AI safety (red-teaming LLMs, blue-teaming with guard models); developing an AI agent framework for finding security vulnerabilities in blackbox, whitebox web applications; generating security consultant reports using retrieval-augmented generation (RAG) and LLM fine-tuning with custom datasets; leading Project Xint Autopen, an offensive security AI engine of Xint, complex usage of LLM agents for automated pentesting and web crawling.
Education
- - KAIST, 2021.9 ~ 2023.8, Daejeon, Korea, M.S. - Artificial Intelligence, under the supervision of Professor Sung Ju Hwang
- - POSTECH, 2016.3 ~ 2021.8, Pohang, Korea, B.S. - Computer Science, GPA 3.70, Major GPA 3.96 (4.3 scale), Magna Cum Laude
Background
- Research interests include machine learning, artificial intelligence, robotics foundational models, vision-language-action (VLA) models, embodied AI, large language models (LLM), and AI safety.
Miscellany
- No specific information provided
