Baseten (Product Review)

My summary of Baseten before using it – Recently, Tuhin Srivastava – Co-Founder at Baseten – talked in a podcast about the Baseten platform. He explained about wanting to help companies operationalise Machine Learning (‘ML’) at scale. Having listened to Srivastava, I’m keen to find out more about Baseten’s value proposition and platform.

How does Baseten explain itself in the first minute? – “From model to full-stack app. Shockingly fast.”

Image Credit: Baseten

To better understand where these claims come from, I scroll down the homepage of the Baseten homepage and see a useful diagram which compares a situation with and without Baseten.

Image Credit: Baseten

The diagram visualises how without Baseten, the process of creating and applying ML models can be very complex, painful and time consuming. Not only for the company using ML but also for the end-user, who will experience a fragmented or slow experience due the different systems an ML model needs to interact with.

Baseten aims to take the pain out this fragmentation, making it quick and easy to get ML models into production. Users can add and deploy a new ML model through a single user interface.

Image Credit: Baseten
Image Credit: Baseten

How does Baseten work? Watching an introductory video by one of Baseten’s engineers, he explains how data scientists can deploy a machine learning model using Baseten’s Python SDK. The deployed model can then be viewed via a dashboard in Baseten, including versions, metrics and logs.

Image Credit: Baseten

Through the dashboard, the machine learning model can be scaled up depending on traffic. Users can modify the model’s code directly in the dashboard too. To deploy a model to a front-end application, users can create an API key and add relevant business logics through so-called “worklets”. A worklet is composed of blocks of different types, like Code, Model and Decision. Each worklet is backed by an API endpoint.

Image Credit: Baseten

Main learning point: Baseten is a platform that aims to make ML application building simpler, quicker and scalable. The way that it combines both the back and front-end aspects of machine learning is compelling and valuable to those businesses that might not have the necessary infrastructure in-house.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: