Snowflake has introduced a serverless fine-tuning service for its Cortex AI platform, enabling data engineers to customize large language models using their own data without managing complex infrastructure. This new feature, Cortex Fine-Tuning, integrates directly into existing data workflows via SQL functions or a no-code UI, bringing the model to the data for enhanced security and governance. The service utilizes parameter-efficient fine-tuning (PEFT) to adapt models from providers like Meta and Mistral AI, making custom LLMs more accessible for specialized enterprise tasks. AI
IMPACT Lowers the barrier for enterprises to create domain-specific LLMs, integrating model customization into data workflows.
RANK_REASON This is a new feature for an existing AI platform, not a frontier model release.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
