Why Your AI Agents Are Guessing, And How Dojo Fixed It With Aiven DataHub

Date and time: 20 October 2026, 16:00 EEST (Finland)
13:00 UTC | 14:00 BST (London) | 15:00 CEST (Berlin) | 09:00 EDT (New York) | 06:00 PDT (San Francisco)
Duration: 45 minutes

AI agents can reach your data tools through MCP, but they can't tell how those tools connect, who owns what, or what "transaction" means in your business. So agents guess.

Dojo, one of the UK's largest payments providers, made Aiven DataHub the context layer for its whole data estate. Join Stan Dmitriev (Aiven) and Elad Leev (Dojo) to learn:

  • How to give agents the right context without flooding them
  • How to find and retire dead dbt models using lineage
  • How to keep metadata accurate with ownership and governance alerts
  • How to connect an agent to DataHub's MCP server (live demo)

You'll leave with a practical checklist for making your data estate agent-ready.

Save your seat
Loading...

Get to know our speakers

Stan Dmitriev

Product Director, Aiven Context

As Product Director at Aiven, Stan Dmitriev solves the "context gap" in modern data systems. Leading the Context initiative, he builds the tools that allow AI agents and users to find, understand, and act on data at scale. With a decade-long track record in product management and startup growth, Stan is a pragmatist who believes the best innovation happens in production. He helps organizations harness open-source data tech to innovate fast without sacrificing reliability.

Elad Leev

Staff Engineer at Dojo

Elad Leev is a Staff Engineer and open-source enthusiast with over 13 years of experience managing complex production systems. He has expertise in distributed systems, databases, and data streaming technologies. Elad is using Kafka in production since 2016, and he is a strong advocate for data streaming solutions, particularly with technologies like Flink, Debezium and Kafka Connect.