Skip to content
View awesome-pro's full-sized avatar
🎯
Focusing
🎯
Focusing
186 followers · 90 following

Achievements

Achievement: Pair Extraordinairex2Achievement: QuickdrawAchievement: Pull Sharkx2Achievement: YOLO

Achievements

Achievement: Pair Extraordinairex2Achievement: QuickdrawAchievement: Pull Sharkx2Achievement: YOLO

Highlights

  • Pro

Block or report awesome-pro

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
awesome-pro/README.md

i do rl post-training on reasoning models & build the inference systems that serve them.

i like working when the world is sleeping. my work cycle is generally 12pm to 4am. And I usually write my thoughts in my artifacts.

i live the part after the launch day. When the reasoning breaks, when the cost skyrockets, when the first traffic hits - all the similar thrills :)

today total

Pinned Loading

  1. hiqcache hiqcache Public

    INT8-quantized hierarchical KV cache for SGLang HiCache. Compresses KV evicted from GPU L1 into the CPU L2 tier, holding 1.78x more tokens in the same host memory. Measured on Qwen3-8B against a BF…

    Python 11

  2. rolloutcore rolloutcore Public

    A versioned RL rollout runtime for vLLM: one weight version per rollout, no cross-version KV reuse, and trajectories bound to the declared training step.

    Python 9

  3. miniserve miniserve Public

    Miniature LLM serving runtime with continuous batching, chunked prefill, KV-cache management, preemption, prefix caching, and real Llama-family execution.

    Python 11

  4. agentflow-pro agentflow-pro Public

    Process-supervised RL for a multi-step reasoning agent - DAPO + a learned Process Reward Model training a Qwen3-8B Planner

    Python 3