Junyi Shen (沈骏一)

Efficient LLM Serving for Agentic Workflows: A Data Systems Perspective

Noppanat Wadlom, Junyi Shen, Yao Lu

SIGMOD 2026 (accepted)

arXiv, code

A workflow-aware serving framework that models agentic workloads as query plans and treats LLM invocations as first-class operators.

Double: Breaking the Acceleration Limit via Double Retrieval Speculative Parallelism

Yuhao Shen, Tianyu Liu, Junyi Shen, Jinyang Wu, Quan Kong, Huan Li, Cong Wang

ACL 2026 (Oral, Best Paper Candidate, SAC Highlight)

arXiv, SAC Highlight certificate

A training-free speculative decoding method that breaks the speedup limit of parallel decoding via double retrieval.

FlowMesh: A Service Fabric for Composable LLM Workflows

Junyi Shen, Noppanat Wadlom, Lingfeng Zhou, Dequan Wang, Xu Miao, Lei Fang, Yao Lu

arXiv preprint

arXiv, code

A multi-tenant service fabric that executes and optimizes these workloads as one shared service instead of isolated pipelines.

Halo: Batch Query Processing and Optimization for Agentic Workflows

Junyi Shen, Noppanat Wadlom, Yao Lu

arXiv preprint (Submitted to The VLDB Journal)

arXiv, code

A system that brings batch query processing and optimization into agentic LLM workflows.

SpecBranch: Speculative Decoding via Hybrid Drafting and Rollback-Aware Branch Parallelism

Yuhao Shen*, Junyi Shen*, Quan Kong, Tianyu Liu, Yao Lu, Cong Wang

ICLR 2026 (Poster)

arXiv

A comprehensive pipeline to accelerate speculative decoding via a hybrid parallelism method.

* Equal contribution.