Democratizing Machine Learning One Step at a Time

At BigML we are working hard to make machine learning easy and accessible. We are happy to see a growing number of customers using BigML to solve very diverse problems from detecting fraud to computing risk to predicting consumer behavior and customer churn.

We believe that one of the keys to fully democratize machine learning is making it very affordable. Many current machine learning solutions are not only costly in terms of complexity, time-to-productivity, configuration, installation, and hardware but also in terms of the licensing costs themselves. We’ve priced BigML using an affordable pre-paid model: with no licensing fees, no monthly fees, and we’ve even thrown in a free development mode. You just pre-pay credits that you can consume as you go. However, we have seen that when users become more familiar with BigML they want to use it to solve many other data-driven tasks, keep their models up-to-date, and create ensembles that may have hundreds of models.

For these users, we’ve been working on a new subscription model that will allow our customers to use BigML as much as they want for a monthly, quarterly, or yearly fee. The only limitations will be the maximum size of tasks and the number of parallel tasks. Customers will be able to select a specific plan or request a personalized one if their needs are different.

BigML subscription plans will be ready very soon. Actually, they are ready but we are still deciding what the right max size, max parallel tasks, and price for each different plan might be. As such, we would like to get  a few users to try our subscription plans for free before setting up the initial parameters. We thought a great way to decide which users should get free subscriptions was by getting them to invite some other friends.

If you invite three friends that sign up for free BigML accounts, you’ll get a free month subscription. With a free subscription you will be able to run unlimited tasks to create unlimited resources (datasets, models, ensembles, evaluations, and predictions). We’ll be limiting size and number of parallel tasks but don’t worry if you feel that our initial limitations are too strong for your current needs we’ll be happy to upgrade them. You’ll be able to get up to a 3-month free subscription plan if you manage to get nine friends to sign up.