engblogs

New research expands limitations of weak supervision, foundation models

An update on the March 20 ChatGPT outage, including findings, actions taken, and technical bug details.

Integrating third-party functionalities with ChatGPT through plugins

Explore the latest research on leveraging foundation models for better performance and specialization.

Exploring the algorithm, data modeling, and system design behind estimating wait times for trip requests at airports to empower drivers strategically.

A case study in improving service mesh performance by addressing Istio propagation delay through investigating xDS lock contention and debounce process optimization.

Introducing ChatGPT plugins that enhance its functionality by providing access to up-to-date information, running computations, and utilizing third-party services while prioritizing safety.

The blogpost discusses how Roblox enables users from all over the world to create and share anything on their platform.

The machine learning behind Airbnb's Categories feature.

This blogpost discusses how Lyft uses graph learning methods to generate embeddings and shares interesting rideshare insights uncovered by embeddings of riders, drivers, locations, and time.

Exploring trends in enterprise ML and the potential of foundation models to change the industry, including examples of companies leveraging them and the risks and complexities involved.

A glimpse into the daily routine of a senior analyst in the Data and Insights Group, supporting the Storytelling and Publishing mission at the New York Times.