Tianyu (Max) Wang

Machine Learning Systems Researcher

I am a Ph.D. student in Computer Science at the University of Pittsburgh, advised by Dr. Longfei Shangguan. My research focuses on efficient systems for generative AI: LLM serving, KV-cache management for video generation, GPU resource sharing, and multi-agent reinforcement learning. I work on the runtime mechanisms that connect model execution, memory reuse, and scheduling. I hold an M.S. in Computational Science from the University of Rochester and bachelor’s degrees in Computer & Information Science and Mathematics from The Ohio State University.

Research projects CV Email

Industry experience

Hewlett Packard Enterprise Labs

May 2025 - Present

Research Intern

Mentors: Dr. Gourav Rattihalli and Dr. Aditya Dhakal

Developing energy-aware workload dispatch and consolidation for serverless LLM serving on shared GPUs.

Enflame Technology Inc.

Jan. 2022 - Aug. 2022

Machine Learning Systems and Frameworks Researcher

Mentor: Dr. Heng Shi

Developed a Rust-based scheduling framework for distributed DNN training and implemented conversions between custom and built-in MLIR dialects.

Publications

Energy-Aware Scheduling for Serverless LLM Serving on Shared GPUs

T. Wang, G. Rattihalli, A. Dhakal, L. Shangguan, D. Milojicic

ATC 2026

WAGES: Workload-Aware GPU Sharing System for Energy-Efficient Serverless LLM Serving

T. Wang, G. Rattihalli, A. Dhakal, X. Tang, D. Milojicic

SC 2025 Workshop

Improving GPU Multi-Tenancy Through Dynamic Multi-Instance GPU Reconfiguration

T. Wang, S. Li, B. Li, Y. Dai, A. Li, G. Yuan, Y. Ding, Y. Zhang, and X. Tang

JSA

A Similarity-based Token Pruning for Vision Transformers in Self-Supervised Learning

S. Li, Q. Tan, Y. Dai, Z. Kong, T. Wang, J. Liu, A. Li, N. Liu, Y. Ding, X. Tang, and G. Yuan

ICLR 2025

STAR: Sub-Entry Sharing-Aware TLB for Multi-Instance GPU

B. Li, Y. Wang, T. Wang, L. Eeckhout, J. Yang, A. Jaleel, and X. Tang

MICRO 2024

Under review

ReelCache: An Efficient KV Cache Compression System for Long Video Generation

T. Wang, D. Yu, Z. Ren, J. Zhang, L. Shangguan

under review at MLSys 2027

Predict, Reuse, and Repair: Accelerating Dynamic Sparse Attention for Long-Context LLM Decoding

T. Wang, G. Rattihalli, A. Dhakal, J. Li, Z. Ren, D. Milojicic, L. Shangguan

under review at ICLR 2027