In Gim 김인

Publications and preprints

• Cacheback: Speculative Decoding With Nothing But Cache
  Zhiyao Ma^†, In Gim^†, and Lin Zhong
  To appear in EMNLP 2025 (Main)
  ^†Equal contributors
• Pie: A Programmable Serving System for Emerging LLM Applications
  In Gim, Zhiyao Ma, Seung-Seob Lee, and Lin Zhong
  To appear in SOSP 2025
• Serve Programs, Not Prompts
  In Gim, and Lin Zhong
  HotOS 2025, May 2025 (pdf)
• Wiretapping LLMs: Network Side-Channel Attacks on Interactive LLM Services
  Mahdi Soleimani, Grace Jia, In Gim, Seung-Seob Lee, Anurag Khandelwal
  Preprint, February 2025 (pdf)
• Asynchronous LLM Function Calling
  In Gim, Seung-Seob Lee, and Lin Zhong
  Preprint, December 2024 (pdf)
• Confidential Prompting: Protecting User Prompts from Cloud LLM Providers
  In Gim^†, Caihua Li^†, and Lin Zhong
  Preprint, September 2024 (code, pdf)
  ^†Equal contributors
• Prompt Cache: Modular Attention Reuse for Low-Latency Inference
  In Gim, Guojun Chen, Seung-Seob Lee, Nikhil Sarda, Anurag Khandelwal, and Lin Zhong
  MLSys 2024, May 2024 (code, pdf)
• Memory-Efficient DNN Training on Mobile Devices
  In Gim and JeongGil Ko
  MobiSys 2022, July 2022 (code, pdf)
• Fast Monte-Carlo Approximation of the Attention Mechanism
  In Gim and JeongGil Ko
  AAAI 2022, February 2022 (code, pdf)