Pane
dcouple/Pane
Terminal-first, agent-agnostic manager for running multiple AI coding agents in parallel — locally or self-hosted.
Software Engineer · AI
I build AI systems that ship.. I take demos to production.. I make agents do real work.. I own it end to end: FE, BE, CI.. I automate what nobody wants to.
Most AI projects die in the demo. I take them the rest of the way — through the frontend, the backend, the pipeline, and the boring parts nobody demos.
- packages published
- 5
- certifications
- 33
- blog posts
- 24
- years shipping
- 7
What I do
Four disciplines, one person to brief.
AI
Agents, tool use, evals and the glue that makes a model useful in production instead of impressive in a notebook.
Frontend
React and TypeScript at scale — the architecture that keeps a codebase fast to work in three years from now.
Backend
APIs, data modelling and the services underneath. Certified across AWS, Azure, GCP and MongoDB.
CI & Automation
Pipelines, releases and the tedious work that should have been a script. GitHub Actions, all the way down.
Open Source
Good code doesn't mind an audience.
Two projects of my own across five published packages, plus fixes merged into repos I don't own — immer, typedash, and Pane, a manager for running AI coding agents in parallel. Good company to keep.
281
journey
rxova/journey
A tiny stepper, wizard and sequencer with zero runtime dependencies.
use-everywhere
rxova/use-everywhere
React hooks for state and messages shared across tabs, windows and workers.
Blog Posts
Blog Posts
I learn it twice — once building it, once explaining it. Architecture, migrations, and the real cost of tech debt.
What are you building?
Bring the part that's stuck. Twenty minutes, no pitch.