Onshoring an agency-built MVP, then making it good
CTO and product lead for a dating startup. I onshored the agency-built MVP, brought the product up to a standard worth scaling, built the engineering team, and grew it internationally.
The problem
Wingman was a dating app built on peer perspective: your friends vouching for you rather than you selling yourself. The idea worked and it had users. But the MVP had been built end to end by an offshore development agency, and that agency still held the app, the backend, the website and the source.
That is a worse position than it sounds. Every change went through someone else’s backlog at someone else’s price. Nobody inside the company could answer a technical question without asking permission first. And the cost of leaving grew every month the arrangement continued.
The product itself was rough. It was an MVP in the honest sense of the term: enough to prove the idea, not enough to build a company on. Reliability, performance and the day-to-day experience all needed serious work before international growth was a sensible thing to attempt.
What I did
I came in as CTO and product lead. The first job was ownership, the second was quality.
- Onshored the agency-built MVP, moving the app, backend, website and source onto infrastructure the company owned and controlled, without interrupting the live product.
- Rebuilt the release and deployment process in-house, which turned shipping from a negotiation into a decision.
- Brought the product up to standard, working through the reliability and experience problems that an MVP is entitled to have and a growing product is not.
- Hired and led the engineering team that took over day-to-day development.
- Scaled the product internationally, with reliability as the explicit priority. A dating product that is down at the wrong moment does not get a second chance.
The outcome
Wingman owned its own technology, could ship independently, and grew into new markets on that footing. It ran that way until the company’s funding ran out in September 2025.
The company not surviving is a separate story from the engineering one. If you are dependent on an agency that holds your source, or sitting on an MVP that has to become a real product, this is ground I have covered before.