DREAM-MoE: Downstream Routing Error-Aware Margin-Preserving Quantization for Mixture-of-Experts Large Language Models
Hancheol Park, Geonho Lee, Tae-Ho Kim. AdaptFM @ ICML, 2026.
Senior AI Research Engineer · Nota Inc.
Hancheol Park
I work on efficient foundation models, MoE-aware quantization, model compression, on-device AI systems, and reliable NLP.
News
2026 only, in chronological order.
-
Published an AWS Technical Blog post on LLM model quantization techniques for AWS Inferentia. Read
-
Won 1st Place in Track C and the Overall Grand Prize at the NVIDIA Nemotron Hackathon Seoul. NVIDIA recap Interview
-
Two MoE quantization papers were accepted to AdaptFM @ ICML 2026. News
About
I am a Senior AI Research Engineer at Nota Inc. My work spans efficient LLM/VLM systems, quantization, pruning, knowledge distillation, model porting, NPU/GPU-aware optimization, vLLM-based serving, and uncertainty-aware NLP. I received my Ph.D. in Computer Science from KAIST, advised by Jong C. Park.
Efficient LLMs/VLMs
MoE Quantization
Model Compression
On-device AI
Reliable NLP
Human-centric CV
Publications
Selected papers. Add local PDFs under papers/ whenever needed.
SRA-MoE: Output-Aware Selective Router Alignment for MoE Quantization
Geonho Lee, Hancheol Park, Seunghyun Lee, Jungwook Choi, Tae-Ho Kim. AdaptFM @ ICML, 2026.
PDF to add
Value-and-Structure Alignment for Routing-Consistent Quantization of Mixture-of-Experts Models
Hancheol Park, Geonho Lee, Tairen Piao, Tae-Ho Kim. arXiv preprint, 2026.
Nota AI at GenAI Detection Task 1: Unseen Language-Aware Detection System for Multilingual Machine-Generated Text
Hancheol Park, Jaeyeon Kim, Geonmin Kim, Tae-Ho Kim. GenAIDetect @ COLING, 2025.
Where do LLMs Encode the Knowledge to Assess the Ambiguity?
Hancheol Park, Geonmin Kim. COLING Industry Track, 2025.
Assessing the Answerability of Queries in Retrieval-Augmented Code Generation
Geonmin Kim, Jaeyeon Kim, Hancheol Park, Wooksu Shin, Tae-Ho Kim. arXiv preprint, 2024.
Self-Knowledge Distillation for Learning Ambiguity
Hancheol Park, Soyeong Jeong, Sukmin Cho, Jong C. Park. arXiv preprint, 2024.
Cluster Self-Refinement for Enhanced Online Multi-Camera People Tracking
Jeongho Kim, Wooksu Shin, Hancheol Park, Donghyuk Choi. AI City Challenge Workshop @ CVPR, 2024.
Road Object Detection Robust to Distorted Objects at the Edge Regions of Images
Wooksu Shin, Donghyuk Choi, Hancheol Park, Jeongho Kim. AI City Challenge Workshop @ CVPR, 2024.
Deep Model Compression Also Helps Models Capture Ambiguity
Hancheol Park, Jong C. Park. ACL Long Paper, 2023.
Question-Answering in a Low-resourced Language: Benchmark Dataset and Models for Tigrinya
Fitsum Gaim, Wonsuk Yang, Hancheol Park, Jong C. Park. ACL Long Paper, 2023. Outstanding Paper Award
Addressing the Occlusion Problem in Multi-Camera People Tracking with Human Pose Estimation
Jeongho Kim*, Wooksu Shin*, Hancheol Park*, Jongwon Baek. AI City Challenge Workshop @ CVPR, 2023. * Equal contribution
Earlier publications
- Hancheol Park, Kyo-Joong Oh, Ho-Jin Choi, Gahgene Gweon. Constructing a Paraphrase Database for Agglutinative Languages. Data & Knowledge Engineering, 2019.
- Huije Lee, Hancheol Park, Wonsuk Yang, Jong C. Park. Detection of Non-Standard Meaning Usage with Word Embedding. HCIK, 2018.
- Wonsuk Yang, Hancheol Park, Jong C. Park. Neural Theorem Prover with Word Embedding for Efficient Automatic Annotation. Journal of KIISE, 2017.
- Hancheol Park, Jung-Ho Kim, Jong C. Park. Addressing Low-Resource Problems in Statistical Machine Translation of Manual Signals in Sign Language. Journal of KIISE, 2017.
- Hancheol Park, Gahgene Gweon, Jeong Heo. Affix Modification-Based Bilingual Pivoting Method for Paraphrase Extraction in Agglutinative Languages. BigComp, 2016. AFNLP Best Asian Paper Award
- Hancheol Park, Gahgene Gweon. Initiating Moderation in Problematic Smartphone Usage Patterns. CHI Extended Abstracts, 2015.
Selected Projects
Sovereign AI Foundation Model Project
Technical owner and lead developer for MoE-specific compression, INT4/NVFP4 quantization, and expert pruning for Solar-Open models.
2025 - PresentLLM Porting and Optimization for Qualcomm NPUs
Optimization and porting workflows for Llama, Qwen, and EXAONE targeting Qualcomm NPU execution environments.
2025Hybrid LLM System for SK Telecom
Hybrid routing system between mobile SLMs and server-side LLMs based on query difficulty, showcased at MWC 2025.
2024Efficient VLMs for On-device Industrial Safety
Lightweight VLMs under 4B parameters deployed on Snapdragon-based mobile and QRB5165 industrial platforms.
2024Awards & Honors
- NVIDIA Nemotron Hackathon Seoul - Track C 1st Place and Overall Winner, 2026.
- GenAI Content Detection Task 1 - 3rd Place, GenAIDetect Workshop, COLING 2025.
- NVIDIA AI City Challenge 2024 Track 4 - 2nd Place out of 53 teams, CVPR 2024.
- NVIDIA AI City Challenge 2024 Track 1 - 3rd Place out of 15 teams, CVPR 2024.
- Outstanding Paper Award - ACL 2023.
- AFNLP Best Asian Paper Award - BigComp 2016.
Experience
Nota Inc.
Senior AI Research Engineer
- Tech Lead, NetsPresso XPU Enabler, Sep. 2025 - Present.
- Team Lead, NetsPresso Application, Sep. 2022 - Dec. 2025.
- Team Lead, NetsPresso Performance, Sep. 2020 - Dec. 2022.
Education
Korea Advanced Institute of Science and Technology (KAIST)
Ph.D. in Computer Science
Thesis: Capturing Ambiguity in Natural Language Understanding Tasks with Information from Internal Layers.
Advisor: Jong C. Park