A niche social media application allows users to connect with sports athletes. As a data engineer, you've designed the architecture of the application to be fully serverless using Amazon API Gateway and AWS Lambda. The backend uses an Amazon DynamoDB table. Some of the star athletes using the application are highly popular, and therefore Amazon DynamoDB has increased the read capacity units (RCUs). Still, the application is experiencing a hot partition problem. What can you do to improve the performance of Amazon DynamoDB and eliminate the hot partition problem without a lot of application refactoring?