Carvago: From zero process to market leader in 8 countries
We became Carvago's technology backbone — CTO as a Service, hybrid teams, and full DevOps — scaling from a single developer to 30+ engineers across 8 European countries.
Carvago.com had an ambitious vision — build the largest online marketplace for used cars in Europe. But when they approached us in April 2020, they had only one in-house developer, zero documentation, and an unfinished codebase. They needed a technology partner who could take full ownership and deliver fast.
„We had positive references for Cookielab, so we approached them as a partner. I'm happy to say they met our expectations. Their team is very professional, knows the technology, and delivers a quality product. It is very difficult to find a software studio that can fully integrate with internal teams and take the project and responsibilities as their own, as Cookielab did. It's important for us to know we can rely on them, and that helps us every day.“
The story
Cookielab stepped in as Carvago's technology partner in April 2020. We provided CTO as a Service, taking full ownership of the technical direction — from architecture decisions to team building and process establishment.
Within the first two weeks, we set up development foundations: migrated repositories, established CI/CD pipelines, implemented SCRUM methodology, and created Jira workflows. We then began hiring and structuring the engineering team, growing it from a single developer to over 30 people.
Our product management team worked alongside Carvago's business stakeholders to gather requirements, validate priorities, and manage expectations. We established equal partnerships with marketing and sales departments, building a customer-oriented culture.
The result: Carvago successfully launched across 8 European countries — Czech Republic, Slovakia, Poland, Italy, Austria, Germany, and more — becoming the fastest-growing automotive marketplace in Central and Eastern Europe.
The challenges
The biggest challenge was time. We had less than 6 months from initial analysis to market launch, starting with zero documentation, unfinished source code, and incomplete infrastructure. The first two weeks were critical — we had to establish all processes while simultaneously preparing for the initial launch.
What's next
Our collaboration with Carvago continues. We're preparing further details about infrastructure improvements — including preparing for massive visitor influx during national TV campaigns and ongoing user testing initiatives.
When we inherited the Carvago codebase in April 2020, we found zero documentation, unfinished frontend (Next.js & React), legacy backend (PHP & Symfony 2), and incomplete infrastructure running on Terraform 0.10. Here's how we turned it into a modern, scalable platform serving millions of users across Europe.
What we delivered
„We chose TypeScript, which allows us to safely rewrite larger parts of the application. We added runtime API response validation and introduced standard libraries. The client got modern React with functional components and hooks — we reduced the massive root state for faster and safer development of new features. Thanks to regular automated app updates, we gained access to new features that accelerate development and performance for higher traffic.“
Technologies
The story
Our first priority was understanding the existing codebase and establishing development foundations. We migrated from BitBucket to GitLab, setting up proper CI/CD pipelines with review apps. The infrastructure was upgraded from ECS to Kubernetes, and Terraform was updated from the outdated 0.10 version.
On the frontend, we chose TypeScript for safe, incremental refactoring of the existing React codebase. We introduced functional components with hooks, replaced the oversized root state with smaller, focused stores, and implemented standard component libraries. Runtime API response validation was added to catch integration issues early.
The infrastructure team built a reliable AWS setup with Kubernetes orchestration, PostgreSQL for data, Redis for caching, and Elasticsearch for search. We implemented comprehensive monitoring via Datadog and alerting through PagerDuty — critical when the platform needed to handle tens of thousands of concurrent visitors during national TV campaigns.
We established a thorough testing strategy: regression tests for stability, smoke tests for deployments, end-to-end tests for critical user journeys, and unit tests for business logic. This gave the team confidence to ship features daily instead of monthly.
What's next
We continue to evolve the platform — optimizing performance for growing traffic, improving SEO indexing, and preparing the architecture for expansion into additional European markets.