Pouya Mahdi Gholami
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
Distributed×1Architecture×1ML×1
Frequent co-authors
Research Timeline
2026
Observation, Not Prediction: Conversation-Level Disaggregated Scheduling for Agentic Serving
The paper proposes scheduling LLM agent workloads at the conversation level rather than the turn level, significantly reducing latency and improving energy efficiency by transforming unpredictable multi-turn inference into a stable, two-phase process.
Highlighted terms show continued research focus across papers