Scale 3D Sensor FusionAdvanced annotations for LiDAR + RADAR data
Scale ImageComprehensive annotations for images
Scale VideoScalable annotations for video data
Scale TextSophisticated annotations for text-based data
Scale AudioAudio Annotation and Speech Annotation for NLP
Scale MappingThe flexible solution to develop your own maps
Scale NucleusThe mission control for your data
Scale Document AITemplate-free ML document processing
Scale CollectCollect diverse data
Scale SyntheticGenerate synthetic data
Our MissionThe age of machine learning is a new phase of computing.Powered by a dynamic research community, computers can now recognize images and audio, translate languages, generate realistic text, and beat humans at games.Machine learning is likely the most significant technological shift happening between 2010 and 2030. Aside from a few innovative products from large tech companies, ML has not yet made a tremendous impact.
Labeled data is the key bottleneck to the growth of the machine learning industry. In fact, labeled data is even more essential than algorithms.
ImageNet is a repository of 14 million labeled images in more than 20,000 categories.
By 2011, AlexNet, the first modern neural network, was the top performer on the ImageNet leaderboard. This kicked off the deep learning craze.
Data labeling is not only practically important, it is also philosophically important to the field. Machine learning is a form of metaprogramming—the developer doesn't directly write the program; the developer writes a program which itself writes the program.
The developer provides a rough framework for what the program should look like (usually a neural network), and what its goal should be (usually a labeled dataset), and that spits out a program that is nonsensical to humans, but is better than any program a human could ever write.
It's a bit demoralizing—as Andrej Karpathy once tweeted, gradient descent writes better code than you. I'm sorry.
Humans can influence both the framework and the goal. The framework is the machine learning architecture and algorithm.
The goal is the labeled dataset. This is the ceiling for how good a model can ever be. The labeled dataset directly programs the final model.
It gets “compiled” into the model via back propagation. As more and more of the software of the world is written with machine learning software, the amount of data required to power these systems will also grow.
Explore Open Datasets→
What we doMachine learning replaces a specialized resource constraint (software engineers) with a commoditized one (data labelers). This significantly accelerates technological progress on a global scale. The growth rate of software will accelerate over the next decade as more software is created by machine learning.
It is early in the age of machine learning. Deeply impactful technology takes time to gestate.
Google was founded 15 years after the invention of the internet. iPhone was invented 24 years after the invention of the internet. It’s only been 8 years since AlexNet was first launched.
Being early makes our work exciting. We have the opportunity to be profoundly influential.
Building infrastructure is more durable than building applications—Scale will amplify the whole impact of machine learning if we succeed, which will reach further than any single application. In this way, we will bend the curve of technology, and therefore the curve of humanity.
Our mission is critically important, and needs to happen now. By structurally increasing the coefficient of growth, we are enabling a new age of software.
Our Investors
Join us as we accelerate the development of AI applications.
