

LONDON Hackathon: Agents & Knowledge Graphs (LangChain x SurrealDB)
This weekend hackathon brings together LangChain and SurrealDB to explore how agents can move beyond demos by grounding themselves in structured, persistent context.
As agentic systems become more autonomous, their biggest failure mode isn’t model capability - it’s context. Agents struggle with accuracy, consistency, and reasoning when memory is shallow or unstructured. Knowledge graphs provide a powerful way to represent relationships, state, and history, enabling agents to reason over how things connect, not just what text is retrieved.
Over the weekend, you’ll build practical agent projects using the LangChain ecosystem, with a strong emphasis on knowledge graphs and real-world retrieval workflows, using SurrealDB as a unified data layer for graph, document, vector, and temporal data.
You’ll also have plenty of time to connect with both the LangChain and SurrealDB communities in London, collaborate with other builders, and get hands-on support from mentors.
Schedule
Friday
18:00 - Doors open, check-in, pizza & drinks
Hackathon kickoff, team formation & briefing (time TBC)
Saturday
All day - Build, iterate, mentor support (details TBC)
Sunday
Final build, demos & judging (details TBC)
15:30 - Closing, winners & wrap-up
Full challenge details and rules will be shared ahead of the event.
What you’ll work on
Agent workflows built with LangChain
Knowledge graph–driven context and memory
Hybrid retrieval patterns (vector + graph)
Persistent agent state and evolving context
Practical, production-oriented agent use cases
Prizes
Swag and LangSmith credits from LangChain
Swag from SurrealDB
Why attend?
This is a hands-on opportunity to build credible agent projects in a single weekend, learn production-oriented patterns, and explore how structured context improves agent accuracy and reliability. You’ll collaborate with experienced builders, get direct support from the LangChain and SurrealDB teams, and connect with the wider London agent and AI engineering community.