Skip to content
View voidreaming's full-sized avatar
16 followers · 26 following

Achievements

Achievement: Pull SharkAchievement: Quickdraw

Achievements

Achievement: Pull SharkAchievement: Quickdraw

Block or report voidreaming

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please donā€™t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this userā€™s behavior. Learn more about reporting abuse.

Report abuse
voidreaming/README.md

Hi, I'm Shouju Wang

Computer Science Ph.D. student researching trustworthy AI agents

Website ā€¢ Google Scholar ā€¢ LinkedIn ā€¢ Email

I'm a Computer Science Ph.D. student at the University of HawaiŹ»i at Mānoa, advised by Prof. Haopeng Zhang. My research focuses on LLM privacy, contextual integrity, multimodal agents, and agent alignment.

I study how AI agents can make appropriate decisions about sensitive information while remaining useful in realistic environments. Previously, I was a research intern at Microsoft Research Asia and Zoom GenAI, and a research assistant at Duke Kunshan University.

Selected research

  • MPCI-Bench ā€” Multimodal pairwise contextual integrity evaluation for language-model agents.
  • Privacy in Action ā€” Practical privacy mitigation and realistic evaluation for LLM-powered agents. EMNLP 2025 Findings.
  • DECAF-GAD ā€” Fair autoencoders for node-level graph anomaly detection. ECAI 2025 Oral.

Outside research, I enjoy literature, rock music, photography, and exploring new places.

Popular repositories Loading

  1. voidreaming.github.io voidreaming.github.io Public

    github pages for personal website

    HTML

  2. OpenAgentSafety OpenAgentSafety Public

    Forked from Open-Agent-Safety/OpenAgentSafety

    Evaluating Agent Safety in Realistic, High-Risk Simulations

    Python

  3. OpenHands OpenHands Public

    Forked from OpenHands/OpenHands

    šŸ™Œ OpenHands: AI-Driven Development

    Python

  4. claude-code-source-code claude-code-source-code Public

    Forked from sanbuphy/learn-coding-agent

    Claude Code v2.1.88 Source Code

    TypeScript

  5. miles miles Public

    Forked from Rockdu/miles

    Miles is an enterprise-facing reinforcement learning framework for large-scale MoE post-training and production workloads, forked from and co-evolving with slime.

    Python

  6. vime vime Public

    Forked from vllm-project/vime

    An LLM post-training framework with vLLM for RL Scaling

    Python