I'm proving the answer inside my own companies — one founder, an AI team, four products already shipping. Thirteen years of Rails behind it; teams from six to eighty-nine.
Not AI-assisted — AI leading.
For 13 years my job was the turnaround: walk into a broken engineering org, restructure it, deliver. I did it on teams up to 89 people.
Then I turned that playbook on myself — and walked away to test one thesis: can one founder + AI build real companies?
Brainz, my holding, is the answer — in public. Four products live. LexPro, built end-to-end with Claude; the firms priced it 10× what I dared to charge.
The surprise, after all those years: code was never the bottleneck. Judgment was. AI took the typing; the hard calls are still mine.
Not AI-assisted. AI leading.
Cover the team's payroll floor at market rate. Surplus pays everything else — products, growth, the people. When an idea hits, the person who built it becomes its CEO. Profit-share, not equity.
The spine of the solo-plus-AI thesis: a data layer that knows the market, and an agent that acts on it.
Code exists; customers are next. The honest tier — surface area visible, not hyped.
Building the holding. Operators run the products they build. Three recurring income streams under the LLC + a portfolio of products in flight under the SAS.
Founded the company. Built the first version of BrainzLab / Sidekiq in October. Pivot from employee to founder.
Mobility platform across Latin America. Scaled infrastructure and led teams of up to 89 people through hypergrowth.
First CTO role. HR tech for the Colombian market. Built and led the engineering organization.
Built ROEPA's digital infrastructure. On and off across France and Spain. The relationship is still active 10 years later as Rails + DigitalOcean maintenance work.