Volca.ai
First engineer at a seed-stage startup. I built the MVP, app and backend, in three months for a conference demo that helped raise a $5.5M seed round.
Context
Volca.ai started as a mobile referral platform and later moved into automated performance pay and payroll. When I joined, the company was its two co-founders, both on the product side and neither technical.
The problem
They had three months to show a working product at a major industry conference. The MVP had to handle one flow end to end: when a referral text message reached the company’s number, the system had to create the lead in the database and push it to the app, where users could see the new lead, ready to contact, and who referred it.
My role
I was the first engineer hired, and the only one. I built the whole React Native app and the Node.js backend.
Key decisions
- Expo, to build the MVP as fast as possible.
- Node.js and PostgreSQL for the backend.
- Twilio to receive the referral text messages.
- LangChain to read free-form SMS written in natural language, extract the people’s details, and map them to our database.
What was hard
The Twilio and LangChain pipeline. Referral messages arrive as free text, written however people write, and they had to become clean, structured records.
After the MVP
As the company moved from referrals to automated performance pay and payroll, I developed core features for the Expo app and the React web dashboard. I improved performance, fixed critical issues and raised the overall stability of the product, working with the backend and product teams on new workflows like automated spiffs, payroll calculations and compensation transparency.
Outcome
The conference demo was a success and helped the company raise a $5.5M seed round. The team has since grown to about 10 people.