RAISL at Rutgers CS
Research systems for capable, efficient, and trustworthy machine learning.
RAISL is Hongyi Wang's research group at Rutgers Computer Science. We build practical machine learning systems across LLM infrastructure, distributed training, federated learning, optimization, and trustworthy evaluation.
Research themes
What We Build
Our work sits at the boundary of algorithms, systems, and practical deployment.
LLM Infrastructure
Training, serving, transparency, and evaluation infrastructure for large language models under real system constraints.
Efficient ML Systems
Algorithms and systems for efficient model training, compression, low-rank adaptation, and distributed optimization.
Federated and Private ML
Learning systems that work across distributed, heterogeneous, and sensitive data while respecting privacy and deployment limits.
Trustworthy ML Workflows
Data, benchmarks, and agentic workflows that make machine learning systems more inspectable, useful, and reliable.
People
RAISL Members
Students and collaborators working on efficient, scalable, and trustworthy ML systems.
Principal Investigator
PhD Students
Undergraduate Researchers
Student Researchers
Join us
Working With RAISL
We welcome students who enjoy building real systems, asking careful research questions, and turning ideas into working artifacts.
Prospective PhD Students
Apply through Rutgers Computer Science and mention RAISL or Hongyi Wang in your application when your interests align with our research areas.
Rutgers MS and Undergraduate Students
Email a short note with your research interests, relevant coursework or projects, and a CV or resume.
Collaborators
We are interested in collaborations around efficient training, LLM infrastructure, open models, federated learning, and trustworthy evaluation.