Skip to content
For Self-Managed AI Platform users running v11.0, see the on-premise platform documentation
Get started > First time here? > Foundational apps

Foundational apps

While getting started with AI is relatively straightforward, scaling these applications and ensuring they meet production standards introduces significant challenges. User authentication, data governance, and model observability must be addressed. This is where DataRobot's Foundational apps come into play, providing a comprehensive framework that simplifies the path to deploying reliable AI solutions.

Foundational apps simplify the process of scaling AI from local experiments to production systems by leveraging industry-standard tools, including Git, GitHub Actions/GitLab Pipelines, and Terraform/Pulumi to provision production-grade resources. In the DataRobot platform, you can manage the complexities of modern AI applications while ensuring that governance, security, and observability are prioritized.

The challenge of scaling AI

Initially, AI tools are user-friendly and provide a great starting point for developers and data scientists. However, as your projects evolve, you encounter a range of complexities. These include ensuring sensitive data is handled securely, implementing user and API authentication to share with colleagues, scaling to handle hundreds of users, establishing governance controls, and maintaining observability over model performance.

In many cases, significant amounts of time and resources are required to build out these components from scratch. A structured approach becomes necessary to not only meet these challenges but also to promote collaboration among colleagues.

DataRobot's solution

DataRobot addresses these scaling challenges by integrating industry-standard, infrastructure-as-code technologies, such as Terraform and Pulumi, into its Foundational apps. These applications allow organizations to easily manage multiple services and components while controlling their lifecycle—all while plugging into standard Git workflows and continuous integration and delivery solutions, like GitHub Actions and GitLab Pipelines.

The following application templates are considered Foundational apps:

Key features:

  • Open Source foundations: There are several open-source Foundational app Templates on GitHub in the DataRobot Community Repositories along with the Declarative API provider via Terraform and Pulumi: DataRobot Community Repositories

  • All-in-one monorepos: A straightforward development environment for all the components of your end-to-end AI solution.

  • Web applications with built-in authentication: Easily create web applications that incorporate robust user and API authentication protocols, ensuring that access is both secure and user-friendly.

  • LLM Deployments with guardrails: As organizations grow, the complexity of managing large language models (LLMs) can increase significantly. Foundational apps provide governance and oversight that aid in the ethical deployment of LLMs.

  • Playgrounds for experimentation: Experiment and test LLM models in isolated environments, allowing for rapid iteration and better learning outcomes without impacting production systems.

  • Use Case tracking: Keep track of multiple Use Cases and their progress, enabling better project management and team collaboration as you scale up your AI efforts.

  • Notebooks and Codespaces: Facilitate training, experimentation, and development with integrated notebooks and coding environments, which help accelerate the learning curve and enhance productivity.