Kobi Bar Hanin.
Building context
for AI agents.

Agents are only as good as the context they reason on.

I spend my days on the hard part of agentic AI: the layer between raw, fragmented company data and an agent that can reason about the business. In practice that means building context graphs that connect warehouses, BI tools, documents and the unwritten rules of how a company works.

I like small teams, fast feedback loops and building our own tools when the cost to build drops. These days that's most of the time.