Bio

I am a first-year PhD student advised by Prof. Vamsi Addanki. My research background is in computer architecture and large-scale AI models. My research interest is in ML systems, with a current focus on Prefill-Decode (PD) disaggregation and inference scheduling. I am particularly interested in the efficient deployment of large-scale AI models in large-scale data centers, considering both software and hardware aspects.

Recent Publications

1. ICLR ’26
   

   Verifier-free Test-Time Sampling for Vision-Language-Action Models
   Suhyeok Jang, Dongyoung Kim, Changyeon Kim, Youngsuk Kim, and Jinwoo Shin.
   International Conference on Learning Representations, 2026.
   PDF  Abstract  BibTeX

   Click here to close the dropdown!

   Vision-Language-Action models (VLAs) have demonstrated remarkable performance in robot control. However, they remain fundamentally limited in tasks that require high precision due to their single-inference paradigm. While test-time scaling approaches using external verifiers have shown promise, they require additional training and fail to generalize to unseen conditions. We propose Masking Distribution Guided Selection (MG-Select), a novel test-time scaling framework for VLAs that leverages the model’s internal properties without requiring additional training or external modules. Our approach utilizes KL divergence from a reference action token distribution as a confidence metric for selecting optimal action from multiple candidates. We introduce a reference distribution generated by the same VLA but with randomly masked states and language conditions as inputs, providing action uncertainty while remaining aligned with the target task distribution. Additionally, we propose a joint training strategy that enables the model to learn both conditional and unconditional distributions by applying dropout to state and language conditions, thereby further improving the quality of the reference distribution. Our experiments demonstrate that MG-Select provides a reliable reference for action selection through task-relevant condition masking and consistently improves base models across diverse simulation and real-world benchmarks.

   Click here to close the dropdown!

   @inproceedings{jang2026verifier,
     title = {Verifier-free Test-Time Sampling for Vision-Language-Action Models},
     author = {Jang, Suhyeok and Kim, Dongyoung and Kim, Changyeon and Kim, Youngsuk and Shin, Jinwoo},
     booktitle = {International Conference on Learning Representations},
     volume = {2026},
     conference = {ICLR},
     pages = {139997--140017},
     year = {2026}
   }
   

2. IEEE Access
   

   ProactivePIM: Accelerating Weight-Sharing Embedding Layer With PIM for Scalable Recommendation System
   Youngsuk Kim, Junghwan Lim, Hyuk-Jae Lee, and Chae Eun Rhee.
   IEEE Access, 2025.
   PDF  Abstract  BibTeX

   Click here to close the dropdown!

   Although deep learning-based personalized recommendation systems provide qualified recommendations, they strain data center resources. The main bottleneck is the embedding layer, which is highly memory-intensive due to its sparse, irregular access patterns to embeddings. Recent near-memory processing (NMP) and processing-in-memory (PIM) architectures have addressed these issues by exploiting parallelism within memory. However, as model sizes increase year by year and can exceed server capacity, inference on single-node servers becomes challenging, necessitating the integration of model compression. Various algorithms have been proposed for model size reduction, but they come at the cost of increased memory access and CPU-PIM communication. We present ProactivePIM, a PIM system tailored for weight-sharing algorithms, a family of compression methods that decompose an embedding table into compact subtables, such as QR-trick and TT-Rec. Our analysis shows that embedding layer execution with weight-sharing algorithms increases memory access and incurs CPU-PIM communication. We also find that these algorithms exhibit unique data locality characteristics, which we name intra-GnR locality. ProactivePIM accelerates weight-sharing algorithms by utilizing a heterogeneous HBM-DIMM memory architecture with integration of a two-level PIM system of base-die PIM (bd-PIM) and bank-group PIM (bg-PIM) inside the HBM. To gain further speedup, ProactivePIM prefetches embeddings with high intra-GnR locality into an SRAM cache within bg-PIM and eliminates the CPU-PIM communication through duplication of target subtables across bank groups. With additional optimization techniques, our design effectively accelerates weight-sharing algorithms, achieving 2.22× and 2.15× speedup in QR-trick and TT-Rec, respectively, compared to the baseline architecture

   Click here to close the dropdown!

   @article{kim2025proactivepim,
     author = {Kim, Youngsuk and Lim, Junghwan and Lee, Hyuk-Jae and Rhee, Chae Eun},
     journal = {IEEE Access},
     volume = {14},
     pages = {245--259},
     publisher = {IEEE},
     title = {ProactivePIM: Accelerating Weight-Sharing Embedding Layer With PIM for Scalable Recommendation System},
     year = {2025}
   }