Challenge

Drishya AI uses machine learning and artificial intelligence to help customers optimize their energy operations. To do this, they derive intelligence from different data sources such as high-frequency Industrial IoT (IIOT) time series sensor data (batch and streaming), alarm data, work journal data, and financial data. To increase speed-to-insights and remove the need for manual analysis, Drishya wanted to design a scalable, reliable, secure, and flexible data platform architecture that would allow them to ingest data from any Energy customer and derive meaningful recommendations from their data quickly and sustainably.

Solution

Drishya first participated in a Design Lab with the AWS Data Lab to design a flexible data platform architecture on AWS that is custom-fit to their business need. Drishya then returned for a Build Lab to accelerate solution implementation by building a working prototype of the proposed design. In four days, Drishya built a multi-tenant data lake using Amazon Simple Storage Service (S3), ETL pipelines using AWS Glue, and pipelines to build and deploy their machine learning models using Amazon SageMaker. The insights and intelligence derived are stored in Amazon Timestream and Amazon RDS PostgreSQL for end-user consumption. See the architecture diagram below.

Benefit

The Drishya team left the AWS Data Lab with a clear direction for building their data platform and taking the solution to production. Drishya successfully launched its data platform with batch use cases only three months post-lab and has since seen a rapid progression in terms of efficiency, capacity, and revenue.

Drishya has been able to onboard clients and scale 2X faster. They have also gained flexibility, with the ability to handle different types of clients from different industrial sectors. These benefits form a linked chain leading from one to the other.

“AWS has helped us rapidly build a world-class, scalable, and secure high-frequency time series platform, which is a core asset enabling us to provide quality business solutions and deliver customer value.”

Saumil Sheth, Chief Operating Officer, Drishya AI Labs Inc.

Architecture Diagram

AWS Services Used

• AWS Glue

• AWS Lambda

• Amazon Simple Storage Service (S3)

• Amazon SageMaker

• Amazon Timestream

• Amazon Relational Database Service (RDS)

• Amazon Kinesis

• Amazon Managed Workflows for Apache Airflow (MWAA)

About AWS Data Lab

AWS Data Lab offers accelerated, joint engineering engagements between customers and AWS technical resources to create tangible deliverables that accelerate data and analytics modernization initiatives. During the lab, AWS Data Lab Solutions Architects and AWS service experts support the customer by providing prescriptive architectural guidance, sharing best practices, and removing technical roadblocks. Customers leave the engagement with a prototype that is custom fit to their needs, a path to production, deeper knowledge of AWS services, and new relationships with AWS service experts.

As published on https://aws.amazon.com/aws-data-lab/drishya-aws-data-lab/