engblogs

Efficient stochastic iterative method for optimization on the Random Generalized Stiefel Manifold without full matrix access

Exploring novel approach for binary classification and anomaly detection using equality separation and closing numbers integration into neural network pipelines.

Introducing RLAIF framework for enhancing code generation in lightweight LLMs with API-usage

Exploring the Locality Barrier and Inductive Scratchpad in Transformer Models for Efficient Learning

Efficiently scale MoEs training with PyTorch and Databricks for state-of-the-art LLMs development

A blog post introducing the Vertex Gen AI Evaluation Service to evaluate the impact of LLMs on business outcomes

Optimizing cross-cloud network connectivity through the Google Cloud partnership with Oracle Cloud Infrastructure

Indian language RAG application using Cohere multilingual embeddings and Anthropic Claude 3 on Amazon Bedrock for searching and querying across multiple languages, showcasing the efficiency of multilingual text embeddings in processing unstructured data.

Automate identification of idle endpoints in Amazon SageMaker using Python script and CloudWatch metrics

Create a self-service digital assistant using Amazon Lex and Knowledge Bases for Amazon Bedrock for efficient customer support and automation.

Explore the innovative ways Databricks leverages enterprise AI internally for various functions, showcasing clear wins and experimental approaches across support, software development, and more.

Introducing enhanced stream recovery capabilities for Datastream enhances data replication resilience