Kevin Hsieh

À propos

I am a researcher at Microsoft Research (opens in new tab) in Redmond (opens in new tab). My research interests lie broadly in the intersection of artificial intelligence (AI) and systems/networking. Recently, I am focused on exploring AI for network security and management, federated and decentralized learning, and mitigating data drifts in machine learning. I received my Ph.D. in the department of Electrical and Computer Engineering (opens in new tab) at Carnegie Mellon University (opens in new tab).

Recent projects:

• Secure East-West Traffic with Zero-Trust Networking Principles
  • GNN-based anomaly detection: NetVigil (opens in new tab) ( (opens in new tab)NSDI’24)
  • Natural-language-based network management: NeMoCopilot (opens in new tab) (HotNets’23)
  • Graphs and systems: Complete Communication Graph (opens in new tab) (HotNets’23)
• Federated Learning
  • Challenges in federated learning: Non-IID data (opens in new tab) (ICML’20), FedDrift (opens in new tab) (AISTATS’23)
  • Application in Space FedSpace (opens in new tab)
• Data Drift mitigation
  • In ML systems: Matchmaker (opens in new tab) (MLSys’22)
  • In video analytics systems: Ekya (opens in new tab) (NSDI’22), RECL (opens in new tab) (NSDI’23)