We build agents that companies can deploy for computational research. We built a custom editor (a ground-up replacement for Jupyter and VSCode) for agents to (a) work in notebooks and (b) coordinate with each other. You can fire off an agent to test a new reward function, spin up five to experiment across competing models, or send a Slack to kick off long-running, autonomous model training jobs. We’ve built a multi-agent coordination system that outperforms CC/Cursor at real research tasks, and are starting trials now with top-10 quant funds and AI labs. We raised a $17M seed round led by Audacious and Conviction, with support from Bloomberg Beta, ex/ante, Wes McKinney (creator of Pandas), and several unicorn founders. Our team is composed of quants (ex-HRT, Citadel), PhD’s, and great founding engineers.