Ram Vegiraju

Master advanced fine-tuning techniques to customize Amazon Nova in this engaging 60-minute hands-on session. Learn to adapt Amazon Nova for financial advisory and investment analysis, including analyzing market data, assessing investment opportunities, and generating personalized financial recommendations.

This hands-on workshop empowers you to revolutionize your personalization use cases by mastering the art of fine-tuning models using Parameter Efficient Fine-Tuning (PEFT). Dive into open source techniques like S-LoRA, LoRAX, and Punica to streamline the training and hosting of high-performance, scalable LoRA adapters.

This code talk uses Amazon SageMaker Studio to demonstrate the complete ML development process, from data preparation and model training to deployment and model monitoring.

In this chalk talk, discover how you can optimize ML models in production with an Amazon SageMaker real-time endpoint. Walk through ways to load test the endpoint for scale, throughput, latency, and cost.

Many customers have ML applications with intermittent or unpredictable traffic patterns. Rather than provision for peak capacity up front, which can result in idle capacity or the need to build complex workflows to shut down idle instances, you can now use Amazon SageMaker serverless inference.

In this technical chalk talk for data scientists and engineering leaders, discover how to deploy end-to-end ML pipelines with Amazon SageMaker for programmatic mediation in real-time advertising. Dive into a customer example to review how SageMaker multi-model endpoints are used to prepare, build, train, tune, deploy, and manage ML models for in-application header bidding with traffic volumes...