Tech Mahindra’s GAiA Platform Provides AI Marketplace and Model Lifecycle Management

R. Bhattacharyya

Summary Bullets:

• GAiA can be deployed on Google, AWS, or Azure clouds, or in a private cloud, on-premises data center, or in a bare metal environment.

• Customers can go to the GAiA public marketplace and download models developed by Tech Mahindra and others, and retune and retrain them for their own use.

Leveraging artificial intelligence to obtain better insights, make more informed predictions, and improve operations is a top priority for most mid-sized and large organizations. However, the process can be daunting. Developing models is time-consuming and skilled resources are scarce and expensive, leaving many organizations in search of solutions that will help them streamline the process. AI platforms strive to do just that by providing a comprehensive environment for developing, training, testing, sharing, deploying, and managing AI models.

Tech Mahindra’s AI platform, GAiA, is the enterprise edition of the open source Acumos platform and is based on its latest Boreas release, which was launched at the end of June 2019. GAiA is designed to make it easier for enterprises to build and manage models, and to collaborate with internal and external stakeholders and partners when deploying AI applications. Not only does it provide tools for developing and training models, it also operates a marketplace in which users can share and download AI models to retrain with their own data. Making the platform even more compelling is the fact that it is cloud and infrastructure agnostic; GAiA can be deployed on Google, AWS, or Azure clouds, or in a private cloud, on-premises data center, or in a bare metal environment. Unlike other model management platforms that are tied to a specific cloud, GAiA can be hosted in the environment of the customer’s choosing. This speaks to the growing preference of customers to work with multiple cloud providers or hybrid environments.

Designed to be an end-to-end platform and ready to use out of the box, GAiA provides tools and capabilities that streamline the adoption of artificial intelligence. Customers can go to the GAiA public marketplace and download models developed by Tech Mahindra and others, and retune and retrain them for their own use. They can then upload the models to an internal marketplace and share them with specific users, another department, across a business, or with other organizations through federation. They can also use the model training and test harness capabilities of the platform before downloading and deploying models into a production environment.

Models are downloaded to the GAiA platform as Docker images that can be retrained, retuned, and reused. Since developers aren’t starting from scratch, solution deployment is accelerated. Functions can be moved to business users so there’s less reliance on savvy enterprise developers and data scientists. Additionally, GAiA’s Design Studio allows users to easily drag and drop models to combine machine learning solutions to solve complex business problems. Furthermore, Tech Mahindra differentiates its platform by complementing it with a suite of professional services, which include model and use case development, model deployment and integration; model customization, retraining, and validation; platform monitoring, support, and features customization; adapters and APIs development; reporting; and infrastructure setup.

Tech Mahindra’s GAiA represents a vision of an end-to-end AI model lifecycle management platform; one that is flexible and open, promotes collaboration, and encourages a broader AI ecosystem.

 

What do you think?

Please log in using one of these methods to post your comment:

WordPress.com Logo

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

Google photo

You are commenting using your Google 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.