Keep ownership clear.
Choose the source of truth. Treat indexes and summaries as views that can be rebuilt.
THE STUDIO / THE APPROACH
Independent thinking.
Inspectable systems.
RecursiveIntell is a founder-led applied R&D studio and public engineering portfolio. Josh Stevenson builds the underlying pieces of useful AI, from agent execution to local memory.
ONE CONNECTED BODY OF WORK
The pieces matter.
So do the seams.
An agent runtime needs to know what it can do. A memory system needs to preserve where its knowledge came from. A tool needs to report what happened when a call failed.
The projects here investigate those questions through source code, focused tests, and public documentation. They range from active software to experimental components; each work page names the scope of its evidence.
Meet Josh Stevenson ↗ENGINEERING PRINCIPLES
Less mystery.
More understanding.
Useful systems make their state, ownership, and failure behavior understandable to the person who has to operate them.
Preserve the boundary.
A local system and an external provider have different responsibilities. Make that visible.
Let evidence travel.
Attach the source, test result, and limitations to the work so another engineer can assess it.
Good systems start
with a clear question.
Hiring for agent infrastructure? Working through a difficult system? Tell me what you’re trying to make work.
Start a conversation ↗