Publications¶
NEO: Saving GPU Memory Crisis with CPU Offloading for Online LLM Inference
Xuanlin Jiang, Yang Zhou, Shiyi Cao, Ion Stoica, and Minlan Yu
Eighth Conference on Machine Learning and Systems (MLSys), May 2025
Don't Stop Me Now: Embedding Based Scheduling for LLMs
Rana Shahout, Eran Malach, Chunwei Liu, Weifan Jiang, Minlan Yu, and Michael Mitzenmacher
International Conference on Learning Representations (ICLR), April 2025
SkipPredict: When to Invest in Predictions for Scheduling
R. Shahout and M. Mitzenmacher
38th Conference on Neural Information Processing Systems (NeurIPS), 2024
Optimal and Approximate Adaptive Stochastic Quantization
R. Basat, Y. Ben-Itzhak, S. Vargaftik, and M. Mitzenmacher
38th Conference on Neural Information Processing Systems (NeurIPS), 2024
Beyond Throughput and Compression Ratios: Towards High End-to-end Utility of Gradient Compression
W. Han, S. Vargaftik, M. Mitzenmacher, B. Karp, and R. Basat
23rd ACM Workshop on Hot Topics in Networks (HotNets), pp. 186-194, 2024
Teal: Learning-Accelerated Optimization of WAN Traffic Engineering
Zhiying Xu, Francis Yan, Rachee Singh, Justin Chiu, Alexander Rush, and Minlan Yu
ACM SIGCOMM, August 2023
SwitchV: Automated SDN Switch Validation with P4 Models
Kinan Dak Albab, Steffen Smolka, Jonathan Dilorenzo, Ali Kheradmand, Konstantin Weitz, Stefan Heule, Minlan Yu, Jiaqi Gao, and Muhammad Tirmazi
ACM SIGCOMM, August 2022
Hashing Design in Modern Networks: Challenges and Mitigation Techniques
Yunhong Xu, Keqiang He, Rui Wang, Minlan Yu, Nick Duffield, Hassan Wassel, Shidong Zhang, Leon Poutievski, Junlan Zhou, and Amin Vahdat
USENIX Annual Technical Conference (ATC), July 2022
Direct Telemetry Access
Jonatan Langlet, Ran Ben Basat, Sivaram Ramanathan, Gabriele Oliaro, Michael Mitzenmacher, Minlan Yu, and Gianni Antichi
ACM SIGCOMM, 2021
Zero-cpu collection with direct telemetry access
Jonatan Langlet, Ran Ben Basat, Sivaram Ramanathan, Gabriele Oliaro, Michael Mitzenmacher, Minlan Yu, and Gianni Antichi
ACM Workshop on Hot Topics in Networks (HotNets), 2021