Avinash Joshi

Join this session to learn how Snowflake ML enables teams to deploy multimodal models — from Hugging Face or custom builds — directly within Snowflake. By natively supporting image, audio and text, organizations can run large-scale inference on unstructured data without managing complex infrastructure.

Monitoring your models in production is critical for ensuring that models behave as expected and business outcomes are safeguarded when performance degrades. Join this session to learn how to use Snowflake's ML Observability to track key model quality metrics (such as performance or drift) over time, with alerts in place when things go awry. This talk will also show how our native ML lineage feature allows you to quickly understand the entire journey of a model for quick workflow audits and debugging.