Streamlining AI Development and Deployment with Portainer
Portainer helps Machina save time and money with their AI development and deployment.
Key metrics
Business overview
Business Focus
Machina specializes in grant management, case management, and application management solutions.Based in Europe
Key Need For Portainer
A solution to help efficiently deploy and manage AI components across multiple customer instances.Container Platform
DockerIntroduction
Machina AS is a Norwegian software company specializing in grant management, case management, and application management solutions. With a growing focus on AI integration, Machina AS faced the challenge of efficiently deploying and managing their AI components across multiple customer instances. This case study explores how Portainer has helped Machina streamline their development and deployment processes, leading to significant time savings and improved productivity.
Background
Machina AS offers a highly customizable, off-the-shelf software platform that allows both the company and their customers to build workflows for various processes. As they expanded into AI integration, they needed a solution to manage containerized deployments of their in-house AI components. With a technical team of 20 developers, AI engineers, and a DevOps team, Machina AS turned to Portainer to address their deployment challenges.
Critical business issues
"Our plan is to use Portainer, especially the CI and GitOps features, to automatically deploy new versions as they get built. And have quicker testing because we can deploy new versions immediately and rollback if needed."
Addressing the Challenge
1. Streamlined deployments and time savings
Before implementing Portainer, Machina AS's DevOps team had to manually deploy updates to their AI containers. This process was time-consuming and prone to human error. With Portainer's automated deployment capabilities, particularly its GitOps features, Machina AS has significantly streamlined this process.
Zoe Matre, a member of Machina AS's DevOps team, explained:
"Our plan is to use Portainer, especially the CI and GitOps features, to automatically deploy new versions as they get built... and have quicker testing because we can deploy new versions immediately and rollback if needed.”
This automated approach has not only saved time but also improved the reliability of their deployments. The impact on time savings has been dramatic. Zoe Matre provides a striking comparison:
"Per day, I spent an hour or two easily on deployments and now with Portainer and full use, that'll probably go down to at most 5-10 minutes."
This reduction from 1-2 hours to 5-10 minutes per day represents a time saving of up to 95% on deployment tasks.
2. Enhanced developer productivity and AI integration
The impact of Portainer extends beyond the DevOps team to the entire development organization at Machina AS. By simplifying the deployment process, Portainer has empowered developers to take more control over their work. Zoe explains:
"Now the developers can either request and we can deploy and click a button, or they can click and deploy themselves, which of course saves a lot of time for everyone."
This shift not only saves time but also reduces the back-and-forth communication between developers and the DevOps team, streamlining the entire development process.
As Machina AS expands its AI capabilities, Portainer plays a crucial role in managing the deployment of AI models. The company plans to use Portainer to manage Docker containers running on individual VMs for each customer's AI instance. Zoe describes their approach:
"Every customer gets a separate VM for their own instance of the software. And that includes the AI components that we're now bringing in because they can be ordered separately from the main software. Our plan is to run Docker on those VMs and continuously deploy new versions of our AI container that does the various interferences."
This containerized approach, managed through Portainer, allows Machina AS to rapidly iterate on their AI models and deliver updates to customers efficiently.
3. Role-based access control (RBAC)
As Machina AS grows, managing access to their development and deployment infrastructure becomes increasingly important. Portainer's RBAC capabilities, integrated with their existing OAuth system, provide a streamlined approach to access management.
Zoe highlights the importance of this integration:
"We don't want to manage permissions in more places than we absolutely have to. We want a single source and then build from there."
By centralizing access control, Machina AS can maintain security while reducing the administrative overhead of managing multiple systems.
"Now the developers can either request and we can deploy and click a button, or they can click and deploy themselves, which of course saves a lot of time for everyone."
Looking Ahead
Machina's use of Portainer is set to expand as they continue to grow. They plan to use Portainer across their development, testing, and production environments, potentially extending its use to containerize their main application as well.
Zoe outlines their future plans:
"We're thinking about containerizing the application itself. So likely at least us in the DevOps team, AI engineers and developers probably... 20 people [will be using Portainer]."
This expansion of Portainer usage across their organization indicates the value they've found in the platform and their confidence in its ability to support their growth.
Metrics
Portainer has had a significant impact on Machina's operations, particularly in their AI development and deployment processes:
- Time Savings: Reduction in deployment time from 1-2 hours to 5-10 minutes per day (up to 95% time saving, on average 80 min saved per day, or $340k total (Using $50/hr equals $17,350 saved per deployer x 20 deployers)
- Team Impact: Approximately 10 people (4 DevOps, 5-6 AI developers) directly benefiting from Portainer
- Expanded Usage: Plans to extend Portainer usage to approximately 20 team members, including DevOps, AI engineers, and developers
- Customer Growth: Projection of 5-10 more customers per year, potentially increasing to 15 additional with new AI capabilities, all managed through Portainer