One aspect we truly appreciate about Refonte is its capability to comprehensively label the world. We can annotate 2D bounding boxes, 3D bounding boxes, as well as semantic segmentation, even in 3D, allowing us to gain a thorough understanding of various scenarios, including those we may not anticipate at present.

Adrien Gaidon

Machine Learning Lead

Since Refonte had already furnished our perception team with quality annotations, it seemed logical to extend our usage of their platform to address related pipeline challenges like data selection and model performance debugging. Their platform's robust search capabilities and user-friendly tools facilitated a smooth transition, allowing us to leverage our existing library of annotations effortlessly.

Cassandra Ung

Software Engineer

Refonte has streamlined our annotation gathering process at a competitive price point. The user interface is intuitive, and the built-in worker evaluation pipeline and batch options save us time while promoting best practices, ensuring we receive high-quality training data.

Jack Guio

Head of Autonomy

After dedicating years to training for this research, I found it frustrating how much time I was spending solely on annotating data. Collaborating with Refonte Rapid liberated my time, allowing me to focus on the aspects of research that demand my expertise.

Caleb Weinreb

Neuroscience Post-Doc

ML models can achieve peak accuracy when they effectively address edge cases, which may pose challenges, be uncommon, or even hazardous. The Autotag feature in Data Engine: Dataset Management greatly aids us by pinpointing instances of rare scenarios within our dataset, all through a straightforward query. At Nuro, where our goal is to ensure efficient and safe deliveries, tools like Refonte Data Engine: Dataset Management are invaluable. They assist us in curating edge cases that we can utilize to train models that are increasingly accurate and capable.

Olivier Monson

Sr. Manager, Data Operations