Building and Orchestrating Real Production-ready Agentic AI Systems
summary
Shows a concrete Java path from local dev to Kubernetes production using LangChain for J and Quarkus dev UI, with live demo of a car‑rental multi‑agent system that calls parallel cleaning/maintenance/pricing agents via the A2A agent protocol, aggregates results, enforces guardrails and human‑in‑the‑loop approval, and records token usage, latency and execution traces for observability. Uses RAG/vector store for domain data, supports images (multimodal), deploys with ConfigMaps/Secrets to Kubernetes, and runs on an OpenAI model or any inference server.
AI-written summary · about
chapters (16)
- 0:00Enterprise agentic AI systems
- 1:05From chatbots to AI apps
- 2:50Core components
- 4:14Security and guardrails
- 6:02Observability in production
- 7:45Multi-agent workflows
- 9:30Java for enterprise AI
- 11:35Orchestration patterns
- 14:00Live demo: car rental system
- 19:43Code structure with LangChain
- 22:12Observability and topology
- 24:36Deploying to Kubernetes
- 26:22Multimodal inputs
- 27:47Future of enterprise AI
- 28:52Advanced features
- 30:17Resources and closing
related talks
31mBuilding and Orchestrating Production-ready Agentic AI Systems
01h 01mScaling AI with Java: Building Production-Ready LLM Applications
47mCreating AI-infused Apps, the Easy Way 🤙
49mAgentic AI Patterns
27mFrom Chatbots to Agents: Orchestrating Agentic AI on Existing Enterprise Kubernetes Infrastructure.
51mSupercharge Agentic AI Apps: A DevEx Driven Approach to Cloud-Native Scaffolding
02h 37mDeveloping Cloud-Native Java AI applications with DJL and LangChain4J
48mAgentic AI with Quarkus, LangChain4j and vLLM