Snowflake AI Research

Conducting open, foundational research to advance the field of AI and make enterprise AI easy, efficient, and trusted.

Read Our Latest Blogs and Publications

Results

Newest - Oldest

Gen AI

Eval-Guided Optimization of LLM Judges for the RAG Triad

Machine Learning

Machine Learning Models Require the Right Explanation Framework — and It’s Easy to Get Wrong

Gen AI

Benchmarking LLM-as-a-Judge for the RAG Triad Metrics

Gen AI

Snowflake Arctic Embed Joins ArcticTraining: Simple And Scalable Embedding Model Training

view more engineering blogs

Results

Newest - Oldest

LLM Evaluation

GEval: Tool for Debugging NLP Datasets and Models

LLM Evaluation

Named Entity Recognition - Is there a glass ceiling?

LLM Evaluation

From Dataset Recycling to Multi-Property Extraction and Beyond

Embedding

Contract Discovery: Dataset and a Few-Shot Semantic Retrieval Challenge with Competitive Baselines

view more research and publicaitons

Featured Research

Snowflake Arctic LLM

Apache 2.0 license model with top benchmarks in complex enterprise workloads such as SQL and code generation, instruction following and more

Snowflake Arctic training capabilities diagram

Snowflake Arctic inference capabilities diagram

Explore our projects and contributions

Snowflake Arctic modelsAccess Snowflake Arctic LLM and Embed models from Hugging Face

LLM Optimization LibrariesAccess repository of software you can use to optimize inference and fine-tuning of LLMs