tomaz-b.bsky.social
@tomaz-b.bsky.social
Did some exploration over the weekend on how to enhance @neo4j.com Cypher MCP server. Here are my thoughts:

towardsdatascience.com/preventing-c...
Preventing Context Overload: Controlled Neo4j MCP Cypher Responses for LLMs | Towards Data Science
How timeouts, truncation, and result sanitization keep Cypher outputs LLM-ready
towardsdatascience.com
September 7, 2025 at 2:25 PM
Reposted
Struggling to evaluate your LLM agents' complex graph queries? Tomaz Bratanic's new article tackles the lack of robust evaluation for GraphRAG in MCP agentic systems. Learn how to build a benchmark for measuring retrieval quality and multi-step reasoning.
How to Evaluate Graph Retrieval in MCP Agentic Systems | Towards Data Science
A framework for measuring retrieval quality in Model Context Protocol agents.
towardsdatascience.com
July 29, 2025 at 8:19 PM
Everyone’s building agents and spinning up MCP servers. But almost no one is talking about evaluation. In my latest post, I introduce a framework for evaluating graph retrieval in @neo4j.com MCP-based agentic systems, using @langchain.bsky.social

towardsdatascience.com/evaluating-g...
How to Evaluate Graph Retrieval in MCP Agentic Systems | Towards Data Science
A framework for measuring retrieval quality in Model Context Protocol agents.
towardsdatascience.com
July 29, 2025 at 4:49 PM
In my latest @towardsdatascience.com blog, I've used @langchain.bsky.social to show how to structure legal docs as a knowledge graph in @neo4j.com , and then implement a LangGraph agent with supporting benchmark to demonstrate the power of agentic GraphRAG.

towardsdatascience.com/agentic-grap...
Agentic GraphRAG for Commercial Contracts | Towards Data Science
Structuring legal information as a knowledge graph to increase the answer accuracy using a LangGraph agent
towardsdatascience.com
April 3, 2025 at 5:29 AM
Over the past months, we've been exploring text2cypher agentic flows with @llamaindex.bsky.social . The result? A comprehensive repository featuring multiple @neo4j.bsky.social workflows, a benchmarking framework, and an web UI.

www.llamaindex.ai/blog/buildin...
January 15, 2025 at 6:38 PM
Reposted
Text embeddings are great at encoding unstructured text, but they aren’t very good at dealing with structured information and operations such as filtering, sorting, or aggregations.

What do you need? Other tools, such as #KnowledgeGraphs :)

bit.ly/3XMWrjC

#Neo4j #llms
The Limitations of Text Embeddings in RAG Applications
Learn how to overcome the challenges of structured data operations in text embeddings in RAG applications using knowledge graphs.
bit.ly
December 19, 2024 at 2:58 PM
We've added an example how to build an advanced @neo4j.bsky.social text2cypher agent with @langchain.bsky.social 's LangGraph. Would love to hear your thought!

python.langchain.com/docs/tutoria...
Build a Question Answering application over a Graph Database | 🦜️🔗 LangChain
In this guide we'll go over the basic ways to create a Q&A chain over a graph database. These systems will allow us to ask a question about the data in a graph database and get back a natural language...
python.langchain.com
December 11, 2024 at 5:23 PM
I'm learning @llamaindex.bsky.social workflows by implementing an advanced @neo4j.bsky.social text2cypher agent. I'm quite impressed by the ease of implementing various, optionally concurrent, flows.
December 10, 2024 at 12:52 PM
My last blog post on @towardsdatascience.com was about the underlying implementation of constructing knowledge graphs with @langchain.bsky.social 's LLMGraphTransformer!

towardsdatascience.com/building-kno...
Building Knowledge Graphs with LLM Graph Transformer
A deep dive into LangChain’s implementation of graph construction with LLMs
towardsdatascience.com
December 1, 2024 at 9:35 PM