FactSet has significantly accelerated its AI model development by leveraging Databricks AI Runtime. The company's Text-to-Formula (TTF) system, which converts natural language into FactSet Query Language, now benefits from fine-tuned open-source models. This shift from a RAG approach to fine-tuning has reduced end-to-end latency and allows for model retraining every two to three weeks. By integrating this process within Databricks' Lakehouse workflows and utilizing serverless GPU compute, FactSet has reduced the time for model setup, training, and deployment from days to mere hours. AI
IMPACT Accelerates AI development cycles for financial data analysis, enabling faster model updates and deployment.
RANK_REASON Article details how a company uses a specific software product to improve its internal processes.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →