Case Study: Pixability accelerates NLP classification with Snorkel AI

How Pixability uses foundation models to accelerate NLP application development by months

Pixability, a data and technology company that helps advertisers find the right content and audience on YouTube, needed a faster way to accurately categorize billions of videos across a highly granular taxonomy. Its existing NLP model and external data-labeling process were too slow and required too many revision cycles to keep up with the volume and complexity of unstructured video metadata.

Using Snorkel AI’s Snorkel Flow, Pixability applied foundation-model-driven, data-centric workflows to generate training data, refine prompts, and build deployable NLP classifiers. Snorkel AI helped Pixability create more than 400,000 programmatic labels, train a 600+ class multi-label model, and reach over 90% accuracy in weeks instead of months, advancing the product roadmap by several months and enabling more precise ad placement.

Pixability

Jackie Swansburg Paulino

CPO