engblogs

Exploring the past and future of data science through a conversation with former U.S. Chief Data Scientist DJ Patil and Snorkel AI's CEO, covering historical insights and key considerations for the field's development.

Interview with former U.S. Chief Data Scientist DJ Patil on the past and future of data science

Investigating gender bias in Large Language Models (LLMs) and demonstrating biased assumptions about men and women.

A blogpost about self-supervised object goal navigation using in-situ finetuning for a household robot that learns a semantic segmentation model of 3D objects and navigates using those representations.

Improved model accuracy and robustness achieved through Dataset Reinforcement, a strategy utilizing data augmentation and knowledge distillation without additional training cost.

Exploring the challenges faced by Slack's mobile infrastructure team in creating a seamless mobile app.

The blogpost explores the integration of Sentry as the default error monitoring tool for Fly.io, providing information on how to set up and use Sentry for error and performance monitoring in production environments.

This blogpost discusses the use of inverse reinforcement learning (IRL) in Google Maps to determine the best routes for users, taking into account factors like estimated time of arrival, tolls, and user preferences.

Highlights of the accomplishments of the dev interns during the internship season.

Cohere joins a cohort of enterprise AI leaders in making voluntary commitments to manage AI risks at the White House event.

This blogpost explores how Pinterest implemented Ray, an open-source framework for scaling AI and machine learning workloads, to improve dataset iteration speed from days to hours, while improving GPU utilization to over 90%, enabling faster last mile data processing.

Best practices for evaluating RAG applications that leverage the chat and reasoning capabilities of large language models (LLM) for chatbots.