DigitalOcean, a cloud provider focused on AI workloads, launched its GenAI Platform, which compiles third party foundational models to enable developers to build AI agents quickly.
The company launched the foundational model platform at its Deploy developer conference in Austin.
DigitalOcean's GenAI Platform is designed to allow customers to set up agents with their data platforms. The company said the GenAI Platform is designed to reach developers on multiple levels. To that end, DigitalOcean features services like GPU Droplets, which enables customers to spin up Nvidia H100 instances in one and eight GPU configurations.
DigitalOcean highlights the rise of boutique AI cloud providers
GenAI Platform features the following:
- The ability to use third-party foundation models to build chat experiences with document analysis, semantic search and image generation.
- A framework that's model agnostic.
- Workflows to create use case specific agents with enterprise data and context.
- The ability to integrate structured data from databases and APIs.
- Guardrails for reliability and safety.
- Support for private interfaces.
Bratin Saha, Chief Product and Technology Officer at DigitalOcean, said the GenAI Platform is designed to integrate with customers' existing infrastructure and cater toward multiple experience levels.
DigitalOcean said the GenAI Platform is available today. The company said it will add new features to the GenAI Platform including support for URLs as a data source, agent evaluations, auto-indexing for knowledge bases and model fine-tuning.
DigitalOcean has an annual revenue run rate approaching $800 million. Last year, DigitalOcean named Paddy Srinivasan CEO. Srinivasan had been CEO of GoTo and before that was co-founder of Opstera and an executive at Oracle and Microsoft.
Here's a look at DigitalOcean's GenAI Platform:
