Amazon SageMaker Canvas and Snowflake integration | S02 EP15
Join this session to learn about Amazon SageMaker Canvas as a low-code/no-code workspace for ML and generative AI and how you can leverage native integrations between Snowflake and Amazon SageMaker Canvas to build and train ML models and customize generative AI solutions using your Snowflake data - without writing a single line of code.
Tony Mullen
Amazon Employee
Published Apr 18, 2024
Todays show featured Claire from the Amazon SageMaker Canvas team and Matt from Snowflake. They discussed how SageMaker Canvas, a low-code workspace, can be used with data from Snowflake's data platform for AI/ML workloads. SageMaker Canvas covers the full machine learning lifecycle from data preparation to model deployment and includes pre-built models. It aims to make ML accessible for non-experts while integrating with pro-code tools like SageMaker Studio.
Matt demonstrated how to connect SageMaker Canvas to a Snowflake dataset, prepare the data using the Data Wrangler feature, and use generative AI capabilities to gain insights. Claire then showed how to build custom tabular prediction models as well as fine-tune large language models like those from Amazon Bedrock using data from Snowflake. The process is designed to be low-code and user-friendly initially, with the ability to increase complexity and controls for more advanced users.
Common challenges discussed included messy data requiring preparation and users understanding which ML problem type to apply. The collaboration between SageMaker Canvas and Snowflake allows data in Snowflake's data warehouses to be easily leveraged for AI/ML workflows in SageMaker, enabling a smooth end-to-end experience. The integration resonates with customers looking to do data warehousing in Snowflake and AI/ML in SageMaker.
Highlights:
- Low-code experience for non-ML experts
- Fine-tune large language models like Bedrock
- Tabular prediction model building
- Data preparation with Data Wrangler
- Generative AI for data insights
- Ability for complex features for advanced users
- Seamless Snowflake-SageMaker integration
- End-to-end data warehousing to AI/ML
Loading...
Tony Mullen - Senior RDS Specialist Solutions Architect @ AWS
Jonathan Katz - Principal Product Manager - Technical @ AWS & PostgreSQL Core Team Member
Claire O'Brien Rajkumar - Principal Product Manager Sagemaker Low Code ML @ AWS
Matt Marzillo - Senior Partner Engineer @ Snowflake
You can check out our past shows from out community page -https://community.aws/livestreams/lets-talk-about-data
Any opinions in this post are those of the individual author and may not reflect the opinions of AWS.