Complementing LLM with agentic AI: a Quarkus story on AI
summary
Shows a Quarkus app that combines langchain4j LLM calls with a drools rule engine and deploys models on OpenShift; demoed a refund chatbot that uses a state machine to collect customer and flight data, applies rules (refund if delay > 60 minutes, $2 per delayed minute, +10% if age > 65) and returns $528 for a 480‑minute delay for a 70‑year‑old. Uses on‑prem model, temperature 0 to reduce hallucinations and works end‑to‑end.
AI-written summary · about
chapters (7)
- 0:00AI and LLM limitations
- 6:00Symbolic AI and rule engines
- 11:04Machine learning with rules
- 16:29Hallucinations in generative AI
- 24:12Agentic AI and challenges
- 27:19Airline refund policy demo
- 39:51Key takeaways and future
related talks
48mAgentic AI with Quarkus, LangChain4j and vLLM
01h 01mCreate #AI Infused #Java Apps with #LangChain4j
39mLLM Agents Gone Wild (And How to Tame Them with Quarkus)
36mCrafting Intelligent Cloud Native applications with Quarkus
44mCrafting Intelligent Supersonic Subatomic applications with Quarkus
47mCreating AI-infused Apps, the Easy Way 🤙
57mPure Java Enterprise AI/LLM Integration (EAI 2.0)
49mPushing LLMs over the Edge: Exploring the Limits of the Possible