10 papers
ActPlane: Programmable OS-Level Policy Enforcement for Agent Harnesses
Yusheng Zheng, Tianyuan Wu, Quanzhi Fu +6
AI agents increasingly run in production through harnesses, the software around the LLM, including an engine that enforces safety and effectiveness policies, e.g., 'run tests befor…
RollArt: Disaggregated Multi-Task Agentic RL Training at Scale
Wei Gao, Yuheng Zhao, Tianyuan Wu +15
Agentic Reinforcement Learning (RL) trains LLMs through multi-turn interactions with environments, producing workloads that mix compute-bound prefill, bandwidth-bound decoding, CPU…
ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
Wei Gao, Yuheng Zhao, Dilxat Muhtar +13
Agentic reinforcement learning (RL) is reshaping LLM post-training, but end-to-end training time is dominated by compute-intensive, multi-turn rollouts whose resource demand varies…
D-PACE: Dynamic Position-Aware Cross-Entropy for Parallel Speculative Drafting
Tianyu Wu, Yu Yao, Zhenting Qi +7
Speculative decoding accelerates LLM inference by having a small drafter propose tokens that a larger target model verifies in parallel. Recent diffusion-based parallel drafters su…
Crab: A Semantics-Aware Checkpoint/Restore Runtime for Agent Sandboxes
Tianyuan Wu, Chaokun Chang, Lunxi Cao +2
Autonomous agents act through sandboxed containers and microVMs whose state spans filesystems, processes, and runtime artifacts. Checkpoint and restore (C/R) of this state is neede…
RollMux: Phase-Level Multiplexing for Disaggregated RL Post-Training
Tianyuan Wu, Lunxi Cao, Yining Wei +11
Rollout-training disaggregation is emerging as the standard architecture for Reinforcement Learning (RL) post-training, where memory-bound rollout and compute-bound training are ph…