Case Study: Sprinklr reduces machine learning inference costs and latency with Amazon Web Services

Lower Latency and Costs Using AWS Graviton2–Based Instances with Sprinklr

Sprinklr, a cloud-first unified customer experience management platform, needed to lower machine learning inference costs while improving latency and customer experience. Working with Amazon Web Services, the company tested Amazon EC2 Inf1 Instances powered by AWS Inferentia for its real-time workloads that were previously running on GPU-based instances.

Amazon Web Services helped Sprinklr migrate its latency-optimized ML models to Amazon EC2 Inf1 Instances, using AWS Inferentia to deliver faster and more cost-efficient inference. The result was more than 30% lower latency on those workloads, significant cost savings, and faster model deployment, with Sprinklr now able to deploy a model in under two weeks.

Sprinklr

Nitin Goyal

Vice President of Engineering