GoGuardian: Safer schools, empowered teachers, thriving students

Overview

GoGuardian, which powers safe, focused learning for half of U.S. K-12 students, migrated its ML infrastructure to Databricks to manage billions of daily inferences for web filtering, classroom management and harm prevention while maintaining a PII-free, COPPA/FERPA-compliant data environment. Using Delta Lake, Lakeflow, Unity Catalog, MLflow and Databricks Model Serving, GoGuardian achieved up to 50% reduction in machine learning operational costs, 90% operational cost savings with its Delphi website classification model, and a 62% reduction in inappropriate device use among students. AI-driven prioritization also cut the volume of records requiring human review for high-risk content by over 95%, from 1 million to 35,000-45,000.

The challenge

GoGuardian needed to manage 4 to 6 billion ML inferences daily across fragmented, disconnected services spanning data management, ML development, model serving and monitoring, causing operational and cost inefficiencies. The company also needed to maintain strict compliance with COPPA and FERPA through a PII-free, privacy-first data environment, while manual data labeling for sensitive use cases like self-harm detection became a bottleneck slowing development.

The solution

GoGuardian migrated to Databricks, using Delta Lake to bring structure and reliability to their AWS S3 environment. They implemented Databricks Lakeflow for a real-time ingestion pipeline with dbt for SQL-based transformations including hashing logic to remove sensitive student data. Serverless infrastructure eliminated manual cluster management, Unity Catalog provided unified governance and access control, MLflow gave end-to-end control over the ML lifecycle, and Databricks Model Serving scaled models — including their proprietary Delphi website-classification model — to meet real-time demands, all within a PII-free environment for data preparation and transformation.

Reported business value

GoGuardian's operational costs for machine learning services dropped by up to 50% across key use cases, with the Delphi website classification model achieving up to 90% savings compared to previous AWS deployments. Schools using GoGuardian's filtering solutions reported a 62% reduction in inappropriate device use among students. AI-driven prioritization of high-risk content cut the volume of records requiring human review by over 95%, from 1 million records to 35,000-45,000.