Sotabase

Career

· Research Intern, Microsoft Azure Research2025–
· PhD Student, The University of Texas at Austin2021–
· Research Intern, Georgia Institute of Technology2017–2017
· M.Sc in Computer Science, Korea Advanced Institute of Science and Technology (KAIST)2015–2018
· B.Sc in Computer Science, Korea Advanced Institute of Science and Technology (KAIST)2010–2014
· Research Intern, HP Labs
· KAIST17, KAIST
· Research Intern, Meta
· Researcher, Samsung Research On-device Lab

Publications (34)

Mr.BiQ: Post-Training Non-Uniform Quantization based on Minimizing the Reconstruction Error
Computer Vision and Pattern Recognition · 2022
50
cited
FFN-SkipLLM: A Hidden Gem for Autoregressive Decoding with Adaptive Feed Forward Skipping
Conference on Empirical Methods in Natural Language Processing · 2024
25
cited
RingLeader: Efficiently Offloading Intra-Server Orchestration to NICs
Symposium on Networked Systems Design and Implementation · 2023
16
cited
Multi-dimensional Parallel Training of Winograd Layer on Memory-Centric Architecture
Micro · 2018
14
cited
Read-ME: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design
Neural Information Processing Systems · 2024
10
cited
How I learned to stop worrying and love learned OS policies
USENIX Workshop on Hot Topics in Operating Systems · 2025
2
cited
On-the-Fly Adaptive Distillation of Transformer to Dual-State Linear Attention
International Conference on Machine Learning · 2025
2
cited
Dataset Efficient Training with Model Ensembling
2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · 2023
1
cited
Ghost Routing to Enable Oblivious Computation on Memory-centric Networks
International Symposium on Computer Architecture · 2021
1
cited
Improving the Throughput of Diffusion-based Large Language Models via a Training-Free Confidence-Aware Calibration
arXiv.org · 2025
1
cited
Optimizing Transformer Inference with Selective Distillation: Layerwise Conversion to Linear Attention
1
cited
Q-Rater: Non-Convex Optimization for Post-Training Uniform Quantization
arXiv.org · 2021
1
cited
$\textit{Read-ME}$: Refactorizing LLMs as Router-Decoupled Mixture of Experts with System Co-Design
Advances in Neural Information Processing Systems 37 · 2024
Beyond Static Cutoffs: One-Shot Dynamic Thresholding for Diffusion Language Models
arXiv.org · 2025
Lowering the Pre-training Tax for Gradient-based Subset Training: A Lightweight Distributed Pre-Training Toolkit
International Conference on Machine Learning · 2023
Sequential Encryption of Sparse Neural Networks Toward Optimum Representation of Irregular Sparsity
arXiv.org · 2021
Sherlock: Reliable and Efficient Agentic Workflow Execution
arXiv.org · 2025
Sotabase
Yeonju Ro | Researcher Profile | Sotabase | Sotabase