Eluna: An Agentic LLM System for Automating Warehouse Operations with Reasoning and Task Execution
Ning Liu, Kalle Kujanp\"a\"a, Zhaoxuan Zhu, P Aditya Sreekar, Kaiwen Liu, Chuanneng Sun, Jorge Marchena Menendez, Matthew Bales, Tianyu Yang, Shahnawaz Alam, Rose Yu, Baoyuan Liu, Kristina Klinkner, Shervin Malmasi
Why It Matters
What makes this one worth your time
This work is relevant for AI engineers and researchers focused on automating complex procedural tasks in industrial settings, offering a practical solution for enhancing operational efficiency.
Eluna automates warehouse operations using a graph-guided multi-agent system with improved task execution accuracy.
Summary
The paper introduces Eluna, a multi-agent system designed to automate warehouse operations by encoding Standard Operating Procedures (SOPs) as directed acyclic graphs. It uses a novel asymmetric episodic distillation method to improve task execution accuracy and efficiency, achieving high expert agreement in production applications.
Key contributions
- Development of a graph-guided, multi-agent framework for SOP execution.
- Introduction of asymmetric episodic distillation to improve model performance.
- Demonstration of high expert agreement in real-world warehouse applications.
Notable insights
- The use of directed acyclic graphs for encoding SOPs allows for structured task delegation and execution.
- Asymmetric episodic distillation helps in refining models by internalizing corrections without inference-time overhead.
Possible limitations
- Not stated in the abstract
Abstract
arXiv:2607.08960v1 Announce Type: cross Abstract: Warehouse operations are governed by Standard Operating Procedures (SOPs) that encode complex, multi-system decision logic, which must be executed reliably under strict time constraints, yet LLM agents lack mechanisms to enforce procedural compliance and degrade under the context overload full SOP specifications introduce. We present Eluna, a production-deployed agentic system for reliable SOP execution. Eluna is a graph-guided, multi-agent framework that encodes SOPs as directed acyclic graphs with progressive disclosure and delegates independent tasks to parallel sub-agents, each with persistent code execution and live data access. To meet production latency and accuracy needs, we use asymmetric episodic distillation where a strong teacher is improved through episodic error memories, then a smaller student is fine-tuned on the corrected trajectories with memory stripped, internalizing corrections without inference-time overhead. On a 13-task benchmark and two production applications, our fine-tuned models match or exceed their teacher, beat all larger off-the-shelf baselines, and reach 94% expert agreement on the ticket processing application.