Neo4j workshop shows how graph shapes can give agents better lakehouse context
AI Engineer · youtube · 2026-07-23
Why agents need graph-shaped context
The workshop argues that vector search and Text2SQL each give an agent only a partial slice of the data, which can produce confident but wrong answers. The missing ingredient is not a better model or query, but better context.
Three reusable graph shapes on lakehouse data
Zach Blumenfeld demonstrates three graph patterns built with Neo4j on top of lakehouse data:
- Trees: a table of contents that helps an agent understand what exists and where to navigate.
- Communities: theme discovery that surfaces patterns nobody explicitly named.
- Paths and cycles: relationship tracing across entities, documents, and records.
He also shows the graph being served to an agent over MCP, and says the same shapes can be applied across BigQuery, Databricks, or Snowflake. The workshop leaves viewers with code they can reuse on their own data and agents.
More from coding & agent
- Open-source course says the harness, not the model, can push a coding agent into the top 5 — strickvl · 2026-07-23
- TERSE proposes an open semantic state protocol for AI apps and agents — Defiant-Juice-2745 · 2026-07-23
- A Codex Remote meme about debugging an agent while it’s still running — Dimillian · 2026-07-23
- Claude Code orchestration layer routes subtasks to cheap workers and escalates on verified failure — StolenDrinks · 2026-07-23
- Reddit user says Codex is silently downgrading from GPT 5.6 Sol-High to Luna-Low — SweetGirlKatie · 2026-07-23
- GLM-5.2 Vision build bolts on MoonViT with a 49.5M projector — airesearch12 · 2026-07-23