Building reliable AI agents with Eclipse LMOS
summary
Eclipse LMOS is an open source Language Models Operating System that ships LMOS Arc (a Kotlin/JVM DSL) and LMOS ADL (a semi-English Agent Definition Language) and runs as a Spring Boot app; it integrates with existing client libraries so engineers wire APIs while business writes ADLs. Deployed patterns power Deutsche Telekom’s agent rollout (about 4.5 million conversations and ~38% better than vendor bots) and heavy‑machinery diagnostics, and the demos show agents built and iterated successfully in brownfield environments.
AI-written summary · about
chapters (11)
- 0:00Introduction and OCX Overview
- 2:00Sponsorship Opportunities
- 3:00Arun Joseph – Building Reliable AI Agents with Existing Enterprise Assets
- 7:30Real-world Agentic Systems at Deutsche Telekom
- 10:30Operational Intelligence and Agent Use Cases
- 13:20Understanding Agentic Programming Paradigms
- 20:00Enterprise Reality – Brownfield vs Greenfield Challenges
- 25:30The Birth of Eclipse LMOS
- 31:30Inside the LMOS Stack (Arc and ADL)
- 37:00Demo – Defining Agents with LMOS ADL
- 46:30Q&A and Raffle Giveaway
related talks
33mEclipse LMOS&AGC: Agentic Computing for Real-World Enterprise Systems
20mBuilding intelligent, secure, and scalable autonomous systems
36mProduction-Ready AI Agents: From LLMs to Small Language Models
50mLet's build an AI agent
26mFrom APIs to Agents: Enterprise AI using Java, Ollama & MCP
30mLessons From Building and Deploying a Multi-Agent System at Scale: An...
41mBringing stories to life with AI, data streaming and generative agents
40mAgents in Action: Advancing Open Source AI Missions