Today I spent some time setting up an orchestration system for AI coding agents. It looks quite promising — there’ll probably be a lot more to tinker with ahead.

Why try it?

• Lately everyone seems to be using coding agents more and more, each with its own strengths.
• Coordinating multiple sessions/workers by hand gets messy, so I wanted to try an orchestration layer to keep the work organized.
• In the long run, this is also a small step toward optimizing my coding flow.

Today’s experiment

• Got it set up, then ran a few small tasks through it.
• The workflow feels clearer, and the agents seem to “know their jobs” better when orchestrated with tracked state.
• Some parts still feel unfamiliar, but overall it seems worth experimenting with further.

Next steps

• Try more complex workflows (multiple workers, review, CI).
• Hook it into personal projects to see if it actually saves time.
• If it works out, I might standardize it as my default workflow.

Short version: today’s orchestration setup for AI coding agents was well worth trying, and there’s probably plenty more to build and fine-tune ahead.