Shuhong Dai
  • about
  • blog (current)
  • publications
  • A Quick Mental Model for Estimating LLM GPU Memory Use

    Before downloading a large model or spinning up a container, it’s useful to know whether an open-source LLM will actually fit on your GPU.

    5 min read   ·   November 17, 2025

    2025   ·   LLMs   GPU Memory

  • Designing a Maintainable Replay Buffer in RL Systems

    A structured and engineering-focused reflection on replay buffer design in RL, emphasizing clarity, extensibility, and long-term maintainability.

    7 min read   ·   October 21, 2025

    2025   ·   RL   System Design   Data Structures

  • Tracing the Root Cause of Missing GPUs in Docker Containers

    A debugging record of why Docker refused to expose GPUs inside a container even though the host recognized them perfectly, and how every layer of the system contributed a small piece to the failure.

    5 min read   ·   August 20, 2025

    2025   ·   Docker   NVIDIA   CUDA

  • A Measure of Range Compression on Different Board Textures

    A mathematical approach to quantifying how different flop textures compress or expand pre-flop ranges in No-Limit Hold’em.

    5 min read   ·   August 03, 2025

    2025   ·   Poker   Probability   Combinatorics   Game Theory   Information Theory

  • Running dm-control on a Headless Server: A Complete Debugging Log

    A practical record of configuring dm-control with Mujoco on a headless Ubuntu server, covering rendering failures, version mismatches, and the final workable setup.

    5 min read   ·   June 15, 2025

    2025   ·   Reinforcement Learning   Mujoco   dm-control   Rendering

  • Newer
  • 1
  • 2
  • Older
© Copyright 2025 Shuhong Dai.