All projects
Archivedwork

13 Startups in 13 Months

Shipped 13 products end-to-end in 13 months

WeDo.agency2020 – 2021
Agency pivoted; products were client-owned and handed off.
13
Products shipped
€52K/yr
Ad savings for client
PythonReactNext.jsNestJSAngularFirebasePostgreSQLDockerFastAPI

Context

WeDo.agency was a small product consultancy. The model: take a client's idea, validate it fast, ship an MVP, and hand off. In 13 months I worked on 13 separate products.

This was where I learned to be a full-stack developer under real pressure.

What we built

The range was wide:

  • AI admin dashboard for a Silicon Valley startup — a Python + FastAPI backend with GPT-3 integrations, React frontend. Delivered in 6 weeks with a team of 3.
  • Ad automation SaaS — a Python scraper + NestJS API + React frontend that automated Google Ads bid management for an e-commerce client. Saved them €52K/year in agency fees.
  • Real-time dashboards — Firebase Realtime Database + React for live sales tracking.
  • Telegram bots — several product bots using the Telegram Bot API + PostgreSQL for state management.
  • E-commerce platforms — Next.js SSR storefronts with PostgreSQL backends.

What I learned from the pace

Shipping 13 products in 13 months is not about cutting corners — it's about making fast, deliberate decisions about what not to build.

Every project taught a different lesson:

  • The AI dashboard taught me to design prompts as first-class engineering artifacts, not afterthoughts.
  • The ad automation taught me that scraping-based integrations will break and you need monitoring from day one.
  • The Telegram bots taught me that state management in conversational UIs is surprisingly complex.

Team lead experience

On the Silicon Valley project I led a team of 3 developers for the first time. The hardest part wasn't the technical work — it was aligning on what done looked like, and learning to review code in a way that taught rather than just corrected.

What I'd do differently

We reused very little across projects because each used whatever tech the client preferred. In retrospect, building a thin internal starter kit for the most common patterns (auth, email, payments) would have saved days per project.