PREM and AWS Join Forces
PREM and AWS have joined forces to deliver an accelerated Generative AI developer experience on both AWS and the PREM platform. To kick off this milestone, PREM co-organized a GenAI Hackathon with AWS in Barcelona, where AWS's ProServe team collaborated with PREM's team to tackle complex use cases by integrating the PREM platform with AWS services. The two-day event brought together talented developers, innovators, and industry experts from both companies to explore the possibilities of this collaboration.
You might wonder, "Why would PREM and AWS team up?"
The answer is simple: their shared goal is to revolutionize the ML developer experience. This partnership merges PREM's innovative platform capabilities—such as autonomous fine-tuning—with AWS's cloud services, unlocking new possibilities for Gen AI application development. PREM has also seamlessly integrated its platform with AWS Bedrock and S3, offering customers a scalable and efficient solution.
During the hackathon, each team kept the following questions in mind:
What is the customer problem or opportunity?
What is the most important customer benefit?
How do you know what customers need or want?
What does the customer experience look like?
This hackathon proved that when you combine AWS and PREM together, developers can fully unlock the true potential of Small Language Models (SLMs) and AI Agents to meet a customer’s needs.
Here are some of the standout use cases showcased at the hackathon
An AI-driven system that responds to CloudWatch alerts by:
- Optimizing resource allocation
- Enhancing database query performance
- Refining application code
The system generates pull requests based on alert types, addressing issues in order of feasibility.
A custom model that converts CloudFormation to Terraform (a highly requested feature, according to AWS representatives)
AWS pricing calculator for the entire architecture:
- Calculate the cost of running different AWS services and quotas based on an input image of the architecture and project goals
- Generate a high-level architecture using a fine-tuned multimodal model, based on transcriptions of customer-solutions architect conversations
DSaaS (Data Science as a Service): Automatic training and deployment of models (similar to PREM-Factory)
- An automated tool for training and deploying machine learning models, similar to PREM-Factory. Aiming towards making data science capabilities more accessible to teams of all sizes and accelerating model development and deployment without needing extensive infrastructure management.
Automated MLOps deployment from README files
- AI that can automatically fulfill user-based deployment of MLOps infrastructure from a README file.This automation tool helps builders reduce time spent on tedious cloud deployment tasks, enabling them to focus on delivering high-value strategic advice and solutions to their customers.
The hackathon demonstrated what and immense amount of innovation, and we can't wait to see the exciting projects that will emerge from this event.
The Foundations of Our Partnership
The PREM-AWS collaboration opens up a world of possibilities, empowering enterprises to confidently build, deploy, and manage AI applications. This partnership centers on the following key strategies:
- Offering the PREM platform as a Software as a Service (SaaS) on the AWS Marketplace
- Incorporating PREM's proprietary automatic fine-tuning capabilities into Bedrock
- Hosting PREM's proprietary Small Language Models (SLMs) on Amazon Bedrock
- Integrating S3 and high-performing foundation models hosted on Bedrock into the PREM Platform
- Implementing BYOE (Bring Your Own Endpoint) in PREM or Bedrock. BYOE in PREM allows customers to use their own custom domain or subdomain as the entry point for their Bedrock-deployed applications, ensuring consistent access, simplified discovery, and enhanced security.
- The addition of S3 Access Grants that map identities from directories like Active Directory or AWS Identity and Access Management (IAM) Principals to datasets in S3. This feature enables the PREM platform to manage data permissions at scale, automatically granting S3 access to end-users based on their corporate identity—a crucial capability for large enterprises with robust data governance policies.
Together, PREM and AWS have streamlined generative AI integration, making it accessible to businesses of all sizes and technical expertise.
Through AWS Marketplace, companies can quickly implement RAG, fine-tune custom models, and experiment without extensive AI infrastructure knowledge. This collaboration aims to democratize generative AI, enabling businesses to create AI infrastructure without requiring machine learning expertise.
Seamless Integrations with AWS Bedrock and S3: Unlocking AI Potential
PREM has integrated AWS Bedrock and S3 into the developer platform as part of this partnership.
Bedrock, a fully managed foundational model service, powers some high-performing models from AWS's Titan series and AI21 Lab's Jurassic and Jamba series found on the platform.
Meanwhile, the integration with S3 provides a scalable, durable, and highly available object store that can handle massive amounts of data. This enables PREM users to store, manage, and analyze large datasets in their repositories for improved data ingestion for RAG, paving the way for more accurate and tailored responses from custom models.
Explore the web using models found on Bedrock with our AI-powered search engine, Premplexity. Perplexity allows you to delve into new topics and summarize content effortlessly, all while harnessing the power of the PREM Platform. By integrating popular search engines like DuckDuckGo, Premplexity offers a smarter, more secure way to search the web—all without compromising your privacy.
It's clear that PREM and AWS' partnership will continue to drive growth, innovation, and success in the AI community. By providing businesses with a comprehensive, scalable, and secure platform for AI development, PREM, and AWS empower developers and enterprises to push the boundaries of what's possible.
Stay tuned for more exciting developments from PREM and AWS.
Read more
Simone Giacomelli
PremAI
