Ben Snively

In this builders’ session, learn how to develop a comprehensive evaluation framework for the outputs generated by Amazon Q Business. Explore how this framework can help you measure key metrics like truthfulness and completeness, allowing you to quantitatively assess the performance of your generative AI applications.

In this session, learn how the Technology Innovation Institute (TII) built the next generation of generative AI technology. Discover how TII expanded the scope of the modern standard language model by applying know-how in large language models (LLMs) to build cross-disciplinary, cutting-edge expertise.

Amazon Titan foundation models (FMs) are pretrained on large datasets, making them powerful, general-purpose models. You can use them as is or privately customize them with your own data.

U.S. federal, state, and local government agencies spend trillions on social and safety net programs including Social Security, Medicare, and Medicaid, with spending increasing every year. The volume of forms the government handles requires a large workforce to manually review and process paperwork.

Organizations working within highly regulated environments, such as healthcare, government, and financial services, face unique challenges when building data and analytics workloads. Join this session and learn how to overcome these challenges and drive insights from your data while supporting security and compliance efforts.

Machine learning involves more than just training models; you need to source and prepare data, engineer features, select algorithms, train and tune models, and then deploy those models and monitor their performance in production. Learn how to set up an ETL pipeline to analyze customer data using Amazon SageMaker, AWS Glue, and AWS Step Functions.

Market segments like finance, insurance, or manufacturing all have documents and data that are very specific to their business. In this chalk talk, we discuss how you can utilize Amazon Comprehend custom entity recognition and classification, and AutoML features, to build state-of-the-art custom models to extract domain-specific terms and classifiers completely automatically.