Interview with Piero Molino, CEO of Predibase: Low-Code Machine Learning and Trends in LLMs
ai News had the opportunity to chat with Piero Molino, CEO and co-founder of Predibase, during this year’s technology expo. We discussed the significance of low-code solutions in machine learning and current developments in Large Language Models (LLMs).
Simplifying Machine Learning with Predibase
Predibase is a declarative machine learning platform designed to streamline the process of creating and deploying machine learning models. The company’s ultimate goal is to make machine learning accessible to a broader audience, including both experienced organizations and developers new to the field.
The platform empowers in-house teams with expert knowledge to boost their capabilities significantly and cut development time from months to only a few days. Additionally, it caters to developers who want to integrate machine learning into their products but lack the necessary expertise.
Avoiding Extensive Machine Learning Coding
By using Predibase, developers can bypass writing extensive low-level machine learning code and instead work with a simple configuration file – called a YAML file – consisting of just 10 lines outlining the data schema.
General Availability Announcement
At the expo, Predibase announced the general availability of its platform.
Easing Infrastructure Provisioning and Data Integration
A key feature of the platform is its ability to abstract away infrastructure provisioning complexities. Users can easily run training, deployment, and inference jobs on a single CPU machine or scale up to 1000 GPU machines with just a few clicks. The platform also enables seamless integration with various data sources, including warehouses, databases, and object stores, regardless of their structure.
Collaborative Model Development
Molino explained that the platform is designed for teams to work together on developing models. Each model is represented as a configuration with multiple versions, allowing users to analyze differences and compare performance.
The Role of Low-Code in Machine Learning
We then delved into the importance of low-code development in machine learning adoption. Molino stressed that simplifying the process is crucial for wider industry adoption and increased return on investment.
Lowering Entry Barrier for Experimentation
By reducing development time from months to a matter of days, Predibase lowers the entry barrier for organizations to explore new use cases and potentially tap into substantial value.
Rise of Large Language Models
Our conversation touched on the increasing popularity of large language models and their transformative impact on ai and machine learning. Molino acknowledged their immense power but highlighted limitations such as per-query pricing models, slow inference speeds, and concerns about data privacy when using third-party APIs.
Addressing Challenges with Virtual Private Cloud Deployment
To tackle these challenges, Predibase is introducing a mechanism that lets customers deploy their models in a virtual private cloud, ensuring data privacy and providing greater control over the deployment process.
Common Mistakes in Machine Learning
As more businesses explore machine learning for the first time, Molino shared his insights into common pitfalls. He emphasized the importance of comprehending data, use case, and business context before diving into development.
Understanding Data, Use Case, and Business Context
Molino highlighted the common mistake of having unrealistic expectations and a mismatch between what businesses anticipate and what is actually achievable. He advised that companies should fully understand their data and use case, both technically and from a business standpoint.
Democratizing Machine Learning with Predibase
The general availability launch of Predibase’s platform represents an essential milestone in their mission to democratize machine learning. By streamlining the development process, Predibase aims to help organizations and developers alike harness the full potential of machine learning.
You can watch our complete interview with Molino below:
