Scribd’s Human-in-the-Loop LLM Pipeline Classifies Hundreds of Millions of Docs in Two Months
The team used a fast LLM to generate candidate labels and rationales, a judge model to flag disagreements, and SMEs to review only the toughest cases, building a reusable human-in-the-loop loop. This workflow let a three‑person crew classify hundreds of millions of Scribd documents in about two months, cutting costs and scaling accuracy across projects.