CUSTOMER STORIES
How INVISTA Brings Data Science to the Business with Snowflake
Earl Carlisle, Data Science Leader at INVISTA, shares how using Snowpark has helped his team provide the business with model predictions, allowing them to save in both time and cost with data-driven decisions.
Story Highlights
- Centralized data that decision-makers can act on: Snowflake allows INVISTA to bring all data together under one platform to deliver actionable insights.
- Building and developing made easy: Being able to use Python within Snowpark has allowed INVISTA to easily develop with their data in Snowflake and augment with GenAI.
- Deployed ML models in just a few days, compared to months: Executives are now able to view data per cost, per model, per run from ingestion to model prediction, allowing them to make the best decisions for their business.
Video Transcript
This transcript was automatically generated
My name is Earl Carlisle. I'm a data science leader at INVISTA.
My purpose is to really empower my team to provide a better data science product to our business, which allows them to make data driven decisions. INVISTA primarily is a business to business company that specializes in chemical manufacturing. Our raw materials usually petrochemical.
And from there, that's what we take and we, you know, synthesize the fibers and the fabrics as well as the polymers and sort of plastics that people use in their everyday life. What we love most about Snowflake is their new release with DML ops solutions. Snowflake is the sort of that that central figure in our data platform because it allows us to bring everything together. Snowpark has made our lives easier.
It's really native in Python. So if you know the fundamentals of Python, you can do really good work. And it's nice with GenAI because now you can augment that and say, hey. Teach Teach me how to do this in Spark, and then you can just translate that over with a few little bits over into Snowpark to get that parallel processing compute.
So we would just run these massive instances all day, twenty four seven, because we felt like the opportunity cost to restart was too much.
So that's, again, where we've seen Snowflake really save us in both time and in cost. When every company is trying to capture cloud cost, it took us two months to deploy a few models.
And so now we can do those in just a few days, and that's that's powerful for us. One of the things we've never been able to give our executives is per cost, per model, per run. And that's something that Snowflake can give us, is from ingestion to model prediction and from model development as well. Now we can really go and say, hey. These people are finding this valuable because they keep wanting us to run these workloads. They keep using our predictions, and we can start to track that in Snowflake, where we've never been able to do that before.
