Today, we are thrilled to announce the general availability and free trial of Predibase, the first low-code AI platform for developers. Our mission is simple: empower engineering teams everywhere to rapidly build powerful AI-powered applications without needing to learn complex low-level ML frameworks. Using our platform, engineers and data scientists alike can build and finetune custom models leveraging architectures from simple neural networks all the way to Large Language Models (LLM’s).
Challenges Building AI Applications
Every company is looking to build AI-powered applications to gain a competitive edge and unlock a new world of use-cases and possibilities. Unfortunately, many ML tools today are unnecessarily complex, limited to a specific niche, or simply not designed to support cutting-edge, production-grade ML applications.
Building AI apps is not your typical engineering task. The deeper you go, the bigger the challenge.
Why should engineers have to write hundreds of lines of PyTorch code or TensorFlow, figure out how to train efficiently on distributed infrastructure, and hack together an end-to-end production pipeline? There has to be a better way to ship AI-powered applications.
We’ve seen declarative technologies – like dbt and Terraform – radically simplify other engineering workflows. At Predibase, we’re now applying this idea to Machine Learning and bringing the approach of Declarative ML to both startups and enterprises. Pioneered by leading tech companies like Uber, Apple, and Meta, Declarative ML enables developers and data scientists alike to rapidly build, iterate, and deploy state-of-the-art models in a few lines of code.
The Future of AI is Declarative
With Declarative ML, users simply define what they want to predict and the platform automagically figures out how to build an end-to-end ML pipeline. Novice users can leverage best-practice model defaults while experts can finely tune any parameters from encoders to L2 regularization, allowing anyone to build AI-powered applications in hours, not months.
Build any ML model in a few lines of a configuration file using Predibase's declarative approach to machine learning.
Our declarative approach has been adopted by engineering and data science teams at industry innovators like Paradigm, Koo, and Payscale, and large Fortune 500 enterprises in healthcare and financial services. As the proven ML platform for software developers, we’re excited to announce that Predibase platform is now GA and offers a 14-day free trial so that a broader set of organizations can reap the benefits of Declarative ML.
From startups to Fortune 500 enterprises, organizations are adopting declarative ML to shave off months development on AI projects.
There are two ways to try out the Predibase Free Trial:
- Predibase Cloud: Train, analyze, and deploy models on our fully-managed cloud platform. No credit card required.
- Virtual Private Cloud: Deploy Predibase in your virtual private cloud (AWS). Great for highly confidential data. No compute or functionality restrictions.
To learn more about the different options and which one is right for you, visit our docs. If you’re ready to get started, sign-up for your free trial. If you’d like to request a custom demo, you can get in touch with one of our AI solution experts.
Introducing Data Science Copilot and LLM’s
As part of today’s launch, we are also revealing two platform capabilities that are live in Beta.
The first is our Data Science Copilot.
At Predibase, we deeply care about democratizing the ability to effectively build and use ML beyond data science teams. One principle we’ve seen hold true, both in our past experiences and at Predibase, is that Machine Learning is fundamentally an iterative process.
“Very rarely is the first model you build your last”
While Declarative ML allows you to build ML pipelines in just a few lines of code, Data Science Copilot recognizes that iteration is a necessary, and often discounted, part of the ML lifecycle. By intelligently suggesting, explaining, and applying recommendations, Data Science Copilot acts as your personalized ML expert, helping you iterate to a better-performing model even faster.
Guided iteration with modeling recommendations provided by our own LLM-powered Data Science Copilot.
Next are Privately Hosted and Customized LLM’s.
Large Language Models have taken the world by storm. Now, every company is asking themselves how they can best harness the power of LLM’s for their organization.
While general-purpose LLM’s like GPT-4 can solve general tasks, what if you would like to tailor the LLM to your data and custom enterprise workflows? Luckily, Predibase and the declarative approach to ML translates well to the realm of LLM’s.
Predibase provides a managed platform for the entire LLM lifecycle, allowing you to prompt LLM’s, fine-tune them to your ML task, and deploy state-of-the-art LLMs securely within your enterprise infrastructure at a fraction of the cost and effort. Predibase already supports many open-source LLM’s including Alpaca, Bloom, and Vicuna, and we are continually adding support for others on a weekly basis.
Finetune, query and deploy LLMs within your own environment on your own data with Predibase.
Both Data Science Copilot and Private, Customized LLM’s are now available in Beta. Sign up for a free trial or reach out to firstname.lastname@example.org if you have questions or want a custom demo.
Get Started Building AI-Applications on Predibase
The potential opportunity of leveraging ML has never been so monumental. Whether you’re interested in traditional ML, deep learning, or LLM’s, the Predibase platform can help you ship faster. If you’re ready for a single tool to build and deploy powerful AI-powered applications, try out Predibase today!
- Sign-up for your free trial of Predibase
- Join our upcoming webinar From Engineer to ML Engineer with Declarative ML to learn more