How kelvin Slashed Deployment Times by 80% and Enabled Developers to Ship Faster with Qovery
Moving beyond fragile AWS ECS/EKS complexity, kelvin leveraged Qovery to automate reproducible staging environments, reduce developer frustration, and achieve consistent 5-minute deployments.

Context: A Fast-Growing SaaS on the Hunt for a Better PaaS Solution
kelvin is a fast-growing SaaS company that leverages open data and machine learning to generate detailed energy renovation plans for buildings. Their white-label solution serves real estate and construction firms across Europe. As a young startup, kelvin needed to prioritize rapid feature development but their infrastructure was holding them back.
Adrien Siami joined kelvin as a tech lead to help build the product and lay down a scalable engineering foundation. Initially, the team ran a basic AWS ECS (Amazon Elastic Container Service) setup with a single production environment. It was a decent starting point, but as the company scaled, the limitations of this manually managed, fragile infrastructure became impossible to ignore.
{{quote1}}
The Challenge: Overcoming DevOps Overload and Cost Constraints
The team’s initial infrastructure presented several key challenges:
- No Staging Environment: The lack of a pre-production or staging environment meant every new feature was deployed directly to production, a risk they could no longer afford as a live service.
- Time-Consuming Deployments: The manual deployment pipeline was slow and inefficient. For example, the build process for their Ruby application didn't properly cache dependencies, forcing a full re-compilation with every deployment. This could take up to 20 minutes, significantly lengthening their "feedback loop."
- Infrastructure Complexity: As a small team, no one wanted to become a dedicated DevOps engineer. The team needed a solution that would abstract the underlying infrastructure complexities. They considered migrating to EKS (Amazon Elastic Kubernetes Service), but quickly realized this would introduce even more overhead.
- Heroku Wasn't an Option: Adrien had prior experience with PaaS solutions like Heroku, but it was not a viable option for kelvin. The primary reason was that Heroku wouldn’t allow them to utilize their existing AWS credits. Additionally, the platform’s performance and reliability had become a concern since its acquisition by Salesforce.
The Solution: The Power of Kubernetes, Abstracted by Qovery
After researching various PaaS options that were compatible with their AWS credits, Adrien discovered Qovery. The platform's promise to provide the power of Kubernetes with a simplified, developer-friendly interface was exactly what the kelvin team needed.
Here’s how Qovery addressed their specific pain points:
- Automated and Reproducible Environments: The team now has dedicated production and pre-production clusters. Thanks to Qovery's seamless integration with GitHub Actions and its CLI, a simple code merge to the develop branch automatically triggers a deployment to their pre-production environment.
- Cost-Efficient Scaling: kelvin’s pre-production cluster now runs in "Karpenter" mode. This feature leverages cost-effective AWS spot instances, which are automatically managed by Qovery to scale services on demand, without the need for manual node configuration. This provides resilience and significant cost savings.
- Abstracted Kubernetes: Qovery's core value for kelvin is its abstraction layer. The team can benefit from the scalability and resilience of a Kubernetes backend without needing to write or manage complex YAML files. This allows the developers to focus on writing code instead of becoming infrastructure experts.
- High-Quality Support: The kelvin team has a direct Slack channel with Qovery's support team. This has provided a fast, reliable avenue for getting solutions to specific technical challenges, such as versioning their machine learning API.
{{quote2}}
The Results: 80% Faster Deployments, Increased Confidence, and Business Impact
By adopting Qovery, kelvin achieved significant, quantifiable results that directly impacted their business:
- Drastic Reduction in Deployment Time: The time it took to deploy code was reduced from as long as 30 minutes to just 5 minutes. This 80% decrease in deployment time has dramatically shortened their feedback loop and enabled them to iterate at a much faster pace.
- Empowered to Scale with Confidence: The ease of creating and scaling services with Qovery has given the team the confidence to tackle major upcoming projects, such as a massive import of open data. They can now launch hundreds of pods for parallel processing without worrying about orchestrating the underlying infrastructure.
- Increased Developer Focus: With deployments and infrastructure largely automated, Adrien and his team spend minimal time in the Qovery console,a sign that the platform is working transparently in the background. This has freed them to focus on what matters most: developing their core product.
- Accelerated Business Wins: The solution has directly helped kelvin win deals, particularly concerning their white-label feature, thanks to the ease of adding a new domain name (a critical component of this feature).
{{quote3}}
kelvin's journey demonstrates that for a fast-moving, product-focused team, the right infrastructure isn't just about technical capabilities. It's about enabling developers to be their most productive selves. By choosing Qovery, kelvin found a solution that provided the power of a modern stack without the complexity, allowing them to accelerate their growth and confidently plan for the future.
Curious to see how Qovery could streamline your own deployments? Book a demo to see it in action and discover how we can help your team move faster.









Suggested case studies
.webp)


