PhD Candidate in Software Engineering · Beihang University

Jinghao Wang

I am a systems researcher working on cloud computing, GPU scheduling, and resource management for AI workloads.

My current research focuses on performance prediction, cluster scheduling, and resource-aware serving for deep learning and multi-agent LLM workloads.

Selected publications Google Scholar
News

DeepShare has been accepted at IEEE CLUSTER 2026.

Publication details

01 Research

Research Interests

My research studies resource management and performance in cloud and AI infrastructure.

01

LLM Systems

Resource-aware serving for multi-model and multi-agent workloads under strict latency and memory budgets.

  • Agentic workflows
  • KV-cache management
  • Cross-cluster serving
02

GPU Scheduling

Assurance-driven scheduling and resource coordination for deep learning jobs in shared clusters.

  • Multi-tenant clusters
  • Performance assurance
  • Resource efficiency
03

Cloud Systems

Performance prediction and resource estimation for complex cloud-native and microservice applications.

  • Microservices
  • Graph learning
  • Resource prediction

02 Publications

Selected Publications

Conference papers on scheduling, serving, anomaly detection, and resource management.

2026 IEEE CLUSTER
Accepted CCF-B

DeepShare: Assurance-Driven Deep Learning Job Scheduling for Multi-Tenant Clusters

Jinghao Wang, Yihang Zhou, Xiao Zhou, Xinlei Zheng, Xiaoyang Sun, Tianyu Wo, Chunming Hu, Renyu Yang

Assurance-driven scheduling for deep learning jobs sharing multi-tenant cluster resources.

Publication materials will be linked when publicly available.

2026 IEEE ICDCS
Conference paper CCF-B

Maestro: Workload-Aware Cross-Cluster Scheduling for LLM-Based Multi-Agent Systems

Jinghao Wang, Xiao Zhou, Xiaoyang Sun, Yihui Zhang, Yilong Li, Tianyu Wo, Xu Wang, Chunming Hu, Renyu Yang

A hierarchical scheduling system that coordinates memory, placement, and workflow priorities for multi-agent LLM serving.

arXiv Program
2025 IEEE JCC
Conference paper CCF-C

LogAD: A Multi-Feature Fusion Approach for Log Anomaly Detection

Guangzu Wang, Lingzhi Zhang, Jinghao Wang, Tianyu Wo, Xu Wang, Chunming Hu

A multi-feature fusion approach for detecting anomalous behavior in system logs.

DOI DBLP
2024 IEEE JCC
Best Paper Award CCF-C

RESCAPE: A Resource Estimation System for Microservices with Graph Neural Network and Profile Engine

Jinghao Wang, Guangzu Wang, Tianyu Wo, Xu Wang, Renyu Yang

A graph-based prediction framework for estimating heterogeneous resource demand across microservice topologies.

IEEE Xplore PDF
View complete publication record on Google Scholar

03 Research Software

Research Software

Open-source infrastructure for AI training, inference, and resource orchestration.

Open source · Lead maintainer

CRATER

A cloud-native platform for AI training and inference, built for practical experimentation with shared infrastructure.

Kubernetes AI workloads Resource orchestration
Explore on GitHub

04 Experience & Education

Appointments and Education

  1. Education

    PhD in Software Engineering

    Beihang University · Advised by Prof. Chunming Hu

  2. Internship

    AI Infrastructure Intern

    ByteDance · TikTok Lumen AI Infra

    Contributed to the development and operation of an AI training platform, focusing on job submission, scheduling, preemption, and production operations.

  3. Industry

    Software Engineer

    ByteDance

  4. Education

    MS in Computer Science

    Beihang University

  5. Education

    BS in Computer Science

    Chongqing University

05 Academic Service

Academic Service

Journal and conference reviewing in distributed systems, cloud computing, and intelligent systems.

Journal reviewing

IEEE Transactions on Parallel and Distributed Systems

TPDS

Journal reviewing

IEEE Transactions on Cloud Computing

TCC

Conference reviewing

IEEE International Conference on JointCloud Computing

JCC 2026

Conference reviewing

IEEE International Conference on Omni-layer Intelligent Systems

COINS 2026

06 Contact

Contact

For research collaboration or academic inquiries, please contact me by email.

wang_jinghao@buaa.edu.cn GitHub Scholar