Mage launches low-code AI dev tool into general availability

Join online with today’s leading executives at the Data Summit on March 9th. Register here.

The ancient term for magician, magician or sorcerer, is now also the name of a Silicon Valley startup that exhibits some of its own magic.

The Santa Clara, California-based company today unveiled its reward low-code tool for general availability to create AI ranking models for product developers. The company said the year-old table has been in private beta for the past 12 months and is working closely with early-paying customers to make its tool user-friendly, intuitive and easy to use.

After working with hundreds of product developers on Airbnb, CEO and co-founder Tommy Dange saw that those developers knew how AI could be used to improve their product, but they also had to rely on data science resources to implement their ideas. . Nowhere in the world do data scientists come cheap

“People who are working on user-facing features, such as engineers and backend engineers – can code but they don’t go to school for machine learning or AI,” Dange said. “They know exactly what it is and what is useful, but they do not have the skills to do it. Existing solutions are not designed and built for product developers. Therefore, we provide web-based tools that enable individuals to build AI models. Enables to create – especially in the case of ranking use.

“And we’ve seen that there’s a lot of demand for ranking models in products. Let’s say you have a lot of news on your home feed or you have a lot of products that you want to sell. People need rankings to optimize it for their users. And machine learning and AI in general are the right thing to do. ”

Cases of use include increasing user engagement by ranking articles, posts, comments, etc. Increase conversions on your user’s home feed or by showing the user the most relevant products to purchase, Dange said.

Works before the table by connecting existing data sources, such as amplitude or snowflakes. Once the user adds their data, the table will provide guided instructions for clearing and enhancing that data to maximize the model’s performance during training. Once the model completes training, product developers can use its predictions in real-time via API requests, Dange said.

Maj offers free hobby tires. When a developer or company wants to train larger AI models and use more real-time API predictions, they will need to upgrade to a Pro subscription tire.

How AI is implemented

To help technologists, data architects, and software developers learn more about how to use AI, VentureBeat asked Mage CEO Tommy Dang the following questions, who offered these details to our readers:

Venturebeat: What AI and ML tools do you typically use?

Table: Skit-Learn, XGBoost, TensorFlow, SHAP.

Venturebeat: Are you using out-of-the-box models and algorithms – from exaPEle, DataRobot or other sources?

Table: When a product developer uses a table to create a ranking model, the table will create a unique model for their specific use case. We do not use box models, we make them per use case. We use open source algorithms in these models such as linear regression, logistic regression, deep neural network, XGBoost, etc.

Venturebeat: Which cloud service do you primarily use?

Table: We mainly use AWS.

VentureBeat: Do you use the many AI workflow tools that come with that cloud?

Table: We do not use many of their AI workflow tools, they do not suit our needs and do not solve our problems.

Venturebeat: How much do you do yourself?

Table: We use airflow for data pipeline orchestration and host it on an astronomer. We use Spark to process big data and AWS EMR to run those Spark jobs. We mostly use AWS for calculations. We have proprietary pipelines and workflows to prepare data, train and evaluate models, and provide model predictions.

VentureBeat: How are you labeling data for ML and AI workflow? And you can share ballpark estimates on the data you’re processing?

Table: We specialize in training tabular and text data that is already labeled. For unlabelled data, we provide guided instructions to help product developers program their structured data programmatically.

[And we are processing] Billions of data points.

Venturebeat’s mission Transformative Enterprise is about to become a digital town square for technology decision makers to gain knowledge about technology and transactions. Learn more

Similar Posts

Leave a Reply

Your email address will not be published.