Graphs as the Backbone of Scalable Agentic Apps
summary
Compares using haststack pipelines as deployable microservices for indexing and retrieval with Langraph as an orchestrator that only holds state and chooses tools, exposing pipelines via Hooks REST endpoints or MCP and scaling with Kubernetes, embedding and vector DBs handled by the pipeline, components have strict input/output contracts and serialize as YAML, and a revised architecture with a central Langraph supervisor plus executor-only nodes cuts token usage roughly 50% while isolating failures to single microservices.
AI-written summary · about
chapters (7)
- 0:00Introduction and background
- 3:20Agentic apps challenges
- 8:43Graph-based orchestrator concept
- 15:28Building pipelines with Haststack
- 21:09Architecture improvements
- 26:15Final blueprint summary
- 27:38Q&A
related talks
01h 59mBuilding Agentic and Multi-Agent Systems with LangGraph
34mEverything is a graph, including LLM Applications (and that’s handy)
56mOrchestrating Agentic State Machines with LangGraph
01h 03mAgents Everywhere
29mBig Graphs, Tiny Contexts: Dev Tools for Agents
51mOptimizing Complex Workflows with Event-Driven Multi-Agentic Approach
59mOptimizing Complex Workflows with Event - Driven - Multi Agentic Approach
27mBuilding Useful Apps With Small, Local LLMs