Joe Nolte, Allegis Group & Torsten Grabs, Snowflake | Snowflake Summit 2022
summary
Snowpark for Python runs native Python on Snowflake and a staffing-company pipeline moves from VMs into Snowflake, eliminating multi-threading and shrinking the runtime stack from six components to two (a stored procedure plus an ETL scheduler). Python packages and Streamlit host models in-database so notebooks become deployable apps, and hybrid tables promise near-real-time writeback to talent systems to remove an existing one hour delay. The work is in public preview and already simplifies deployment and iteration.
AI-written summary · about
chapters (13)
- 0:00Introduction and guests
- 0:50Allegis Group and role
- 1:52What is Snowpark
- 3:21Simplifying data pipelines
- 4:46Merging app and analytics stacks
- 5:47Streamlit and model delivery
- 6:51Frictionless deployment benefits
- 8:16Speeding up iterations
- 8:58Frontline benefits for staffing
- 10:25Moving dashboards to Snowflake
- 11:28Data platform consolidation
- 12:11Hybrid tables for real-time access
- 13:15Summary and conclusion
