Scale Rapid: Fast Experimentation with Production-Quality Labels

Scale Converge: The Future of eCommerce & Retail with AI

How to Tame the Long Tail in Machine Learning

Scale AI and Research Partner MIT Media Lab Analyze Bias in Dermatological Datasets to Improve Diagnoses

Own Your Own Maps. Introducing: Scale Mapping

OCR Is Not Enough. Try Scale Document AI.

Life at Scale

Doing the Data Work to Mitigate Data Cascades

Why I Joined Scale AI

 Scale AI’s Series E: Deploying AI Across Every Industry

Scale AI Named to CB Insights’ AI 100 2021 List

Scale Transform: AI Leaders Showcase Research Breakthroughs and Real-World Impact

Why I Joined Scale

Getting Started with Nucleus

Making Pytorch Transformer Twice as Fast on Sequence Generation.

Scale AI’s Series D: Expanding our Team to Empower AI Dreamers

Scaling Menu Transcription Tasks with Scale Document

Scale Hackathon

Quick Start Guide to Customized API Documentation

Scale AI Machine Learning Digest - Q3 2020

Introducing: Our New API Docs

Reaffirming Our Commitment to Building Customer Trust: Scale + SOC 2

Welcoming Brad Porter to Scale AI as our first CTO

Discovering Useful Sentence Representations from Large Pre-Trained Language Models

If You’re De-Biasing The Model, It’s Too Late

Introducing: Scale Nucleus

Keeping The Team Close While Distancing

Introducing: PandaSet. The world's most advanced LiDAR dataset for commercial use

How we Scale Machine Learning

Now Available: Scale Document

Introducing a New Look for Scale AI Products

Scale AI Machine Learning Digest - Q1 2020

Supporting researchers and our community during COVID-19

Open Sourcing the World's First AV Dataset for Wintry Environments

How to Label 1M Data Points/Week

Scale AI Selected as One of America's Most Promising AI Companies

Scale AI Selected as One of LinkedIn’s 2019 Top Startups in the U.S.

Is Elon Wrong About LiDAR?

Scale AI’s Series C: Building the data platform to accelerate machine learning

Why We Joined Scale

Introducing Point Cloud Segmentation

Positive Externalities: Life of a Labeler

Training Data - Quantity is no Panacea

Scale Engineering, Product and Design Teams

nuScenes by Aptiv ImageNet for Autonomous Vehicles

Announcing the Scale and Ouster Data Integration Partnership

Metadata to Enrich Image Annotations

Introducing AI-Assisted Semantic Segmentation at Scale

How We Made It Easy for Anyone at Our Company to Trigger Alerts

Announcing Our Series B with Index Ventures and Accel

Adding a Dimension: Annotating 3d Objects with 2d Data

How to make MongoDB not suck for analytics

Scale Engineering Team++

Introducing Teams

Introducing Our New Dashboard

Increasing the Performance of Dynamic Next.JS Websites

Guest Post: Cloudinary Uses Scale to Focus on Image Compression

Announcing Our Series A With Accel

Training Data For Self Driving Cars With Scale

David Garber

Toyota

Oliver Cameron

Voyage

Brandon Moak

Embark

Anantha Kancherla

Lyft

Geoffrey Irving

OpenAI

Rachel Basse

Skydio

Greg Stewart

Skip

Jonathan Hirokawa

Sea Machines Robotics

Virginia Puccio

Standard Cognition

Nolan Lunscher

Otto Motors

Michaela Mann

Optimus Ride

Andy Mahdavi 

States Title

Scale Transform X | Scale AI Conference

The Learning Curve: Scale Rapid

Engineering Women @ Scale: University Recruiting Panel 

The Learning Curve: Scale Mapping 

The Learning Curve: Technical Integrations with Zenseact

Scale Converge | Scale AI Conference

The Learning Curve: Scaling Up Document Processing 

The Learning Curve: Model-in-the-Loop Data Curation

The Learning Curve: Focusing on the Developer Experience

Collision Conference: Building Trust in AI

Scale Transform | Scale AI Conference

MITTR - EmTech Digital: Data is the New Code

The Learning Curve: What's New in Scale AI's Computer Vision APIs

The Learning Curve: Scaling Data Labeling with Machine Learning

The Learning Curve: Fireside Chat with Emily Denton

NeurIPS 2020

The Learning Curve: Fireside Chat with Dr. Yisong Yue

The Learning Curve: Applying Game Rating Systems to Data Labeling

Fireside Chat with Alexandr Wang

The Learning Curve: Debugging Model Performance with Scale Nucleus

The Learning Curve: Tech Talk with Scale AI

The Learning Curve: Humans-in-the-Loop Don't Perform Gradient Descent

NYC Leaders in AI Dinner 

O'Reilly Strata Data & AI Conferences

Tech.AD Berlin

MOVE 2020

Scale AI Office Warming

The AI Summit | New York

NeurIPS 2019

Women’s Trivia Night

Ice cream Social and Game Night

Scale AI Happy Hour

Jeremy Howard is an entrepreneur, business strategist, developer, and educator. Jeremy is a founding researcher at fast.ai, a research institute dedicated to making deep learning more accessible. He is also a faculty member at the University of San Francisco, and is Chief Scientist at doc.ai and platform.ai.

Deep learning researcher & educator

Jeremy Howard

Christian Szegedy is a researcher at Google Brain. He discovered adversarial examples, invented BatchNorm, and his computer vision research has laid the foundations for modern convolutional neural network architectures. Today, he's working on formal reasoning and dreams of creating an automated software engineer.

Staff Research Scientist at Google AI

Christian Szegedy

Scale is trusted by leading ML teams to drive high model performance with the ability to handle long-tail edge cases. Enhance data quality, accerlate iterations on models, increase engineering productivity, and deliver value on AI investment.

The AI Platform for Computer Vision Applications

Scale Mapping is our latest solution to help customers develop and scale their own custom maps. In this tech talk, we discuss the technical and operational challenges inherent in scaling HD maps and share how we scale mapping for our customers.  

Scale Mapping

Large industries like financial services, logistics, healthcare, and government still run on paper documents. At Scale AI, we are building adaptive machine-learning models to scale up document processing without putting the burden on the customer. In this Tech Talk, we discuss the challenges inherent in document processing, and share our development process from designing new datasets to iterating on model performance.

Scaling Up Document Processing 

Zenseact (formerly Zenuity) is a software company focused on developing advanced driver-assistance systems (ADAS) and autonomous driving software. Bolin Shao, Product Owner at Zenseact joins Shaun VanWeelden, Head of Field Engineering Manager at Scale to discuss Zenseact’s journey, the outcomes, and some of the specific tactics developed to have a tight integration with Scale.

Technical Integrations with Zenseact

Scale Nucleus was developed to help ML teams debug their models by debugging their datasets. In this talk, we walk through some of the latest updates and discuss how Nucleus enables you to diagnose model failure cases, to then prioritize what data should be labeled next to improve model performance. We will then walk through a demo of how you can use models to debug data, and data to debug models for model-in-the-loop data curation. 

Model-in-the-Loop Data Curation

Artificial intelligence uses algorithms to make sense of and act upon diverse, complex and fast-moving data — its entire reason for being. CIOs responsible for enabling AI initiatives need to foster a culture of data literacy to drive success with AI-based systems.

This report helps CIOs and business leaders overcome key challenges related to AI including:
- The consequences of getting data "right" or "wrong" and unintentionally reinforcing harmful biases
- Beliefs that AI algorithms and functionality are more important than data in AI-powered applications and platforms
- Bridging the gap betweenbIT and business professionals involved in AI

Artificial Intelligence Demands That CIOs Foster a Data-Literate Society

"By 2020, 50% of organizations will lack sufficient artificial intelligence (AI) and data literacy skills to achieve business value"

Moving from “I want to use AI” to a tactical approach aimed at practically and sustainably solving business problems, while managing expectations, does not require superpowers. CIOs can follow these five steps to pursue an AI strategy in a pragmatic fashion.

5 Steps to Practically Implement AI Techniques

In this webinar, Scale's Field Engineering team will be presenting their work to improve the customer developer experience with the Scale API. Specifically, we will be covering:

- Our new Python SDK, version 2.0 
- A new Data Pipeline Starter Kit
- Scale's interactive documentation
- Scale's Lidar Toolkit to enable fast creation of Sensor Fusion Tasks

Focusing on the Developer Experience

Scale AI CEO and Founder Alex Wang joins CNBC's Julia Boorstin to share how he grew his transformational startup from an MIT dorm room to a multi-billion dollar platform serving companies across every industry -- including financial services, e-commerce, and autonomous vehicles. Alex addresses concerns about bias in AI, calling it "one of the biggest problems that we need to be solving today", and describes how Scale has helped tackle the problem with products like Nucleus. 


Building Trust in AI

Software development used to be “write code, deploy code, evaluate.” Now it’s “label data, train model, evaluate.”  Overnight, labeled data has become the bottleneck to the growth of the machine learning industry. Is it actually possible to use machine learning to label datasets and train models faster?

Data is the New Code

We've been hard at work creating more flexibility and control for your Computer Vision labeling needs. In this deep-dive webinar, Scale AI's Field Engineering Manager, Shaun VanWeelden, will walk through the latest endpoints, parameters, and workflows to take your labeling pipelines to the next level.

What's New in Scale AI's Computer Vision APIs

At Scale AI, we label on the order of 10MM annotations per week. Our data is diverse both in image space (e.g. cameras, weather conditions, driving surface) and label space (e.g. object and attributes categories). In this talk, Felix Lau discusses how we leverage machine learning to scale a customer-specific, problem-specific, real-time data labeling solution.

Scaling Data Labeling with Machine Learning

In this tech talk, held in conjunction with NeurIPS, Russell Kaplan, Head of Nucleus at Scale AI, is joined by Dr. Yisong Yue for a fireside chat. Russell and Dr. Yue will discuss how research that's happening in academia might translate to industry in the future.

Fireside Chat with Dr. Yisong Yue

In this tech talk, held in conjunction with NeurIPS, Aerin Kim, Engineering Manager of Scale AI’s ML Team, is joined by Emily Denton of Google’s Ethical AI team for a fireside chat. Aerin and Emily will discuss the role of subjectivity and interpretation in data annotation and take your questions.

Fireside Chat with Emily Denton

In this tech talk, we show how game rating systems can be applied to NLP and Search relevance data labeling at Scale AI to track the performance of crowd-workers on our platform.
<br>
Agenda:
- The challenge of skill assessment
- Overview of game rating systems
- How game rating systems work
- Analysis of the results, pros & cons, and areas for future improvement

Applying Game Rating Systems to Data Labeling

In this Tech Talk, we show how you can achieve the concept of [“Operation Vacation”](https://youtu.be/oBklltKXtDE?t=508) for the models you create, and make sure that the model is testing the subsets that you actually care about with Scale’s latest product, Scale Nucleus. 
<br>
Using [PandaSet](https://pandaset.org/), a multimodal dataset for autonomous driving, we demonstrate how you can easily debug model performance and automatically refine your model. In the process we also dive into Nucleus’s features to show how to curate sub datasets and edge cases easily with custom metrics, image similarity search, and auto-pivot, automatically augmenting the data collection to accelerate machine learning training process.

Debugging Model Performance with Scale Nucleus

In this Tech Talk exclusive to Microsoft, LinkedIn, and GitHub employees, Robert Aichner, Microsoft, Principal Group Program Manager - IC3 AI and Machine Learning joins [Alexandr Wang](https://www.linkedin.com/in/alexandrwang/), CEO of Scale AI to discuss how he is accelerating the development of AI applications for the Teams suite of products.

Tech Talk with Scale AI

Data labeling at scale is a complicated endeavor involving operations, machine learning and game theory. In this talk we focus on the game theory, and introduce a model of human-in-the-loop labeling as an optimization problem. Based on our (Scale AI’s) experience of operating tens of thousands of labelers on high complexity tasks, we will discuss how various problems predicted by the model map to observed results.

Humans-in-the-Loop Don't Perform Gradient Descent

AI has had monumental impact on enhancing search relevance and ranking with significant advancements in capabilities for text, speech, and classification. But the possibilities for AI are still nascent. Learn how to take NLP applications to market faster in a diverse set of languages.

Search & Relevance

AI has made monumental impact on enhancing video conferencing application, AI Assistants, Chatbots and more. But the possibilities for AI are still nascent. Learn how to accelerate the development of ML-powered applications with high-quality training data by downloading our solution brief.

Speaker Detection & Noise Suppression

Machine learning is a rapidly evolving field with significant advancements being made by a robust research community. To support leading teams working on cutting edge research, we are pleased to offer Scale for AI Research. 
<br>
__What is Scale for AI Research?__
Starting with our NLP product offering ([Scale Text](https://scale.com/text) and [Scale Document](https://scale.com/document)), we will provide research teams with at-cost annotations for the development of open-sourced datasets to push forward AI research. We are especially interested in supporting ethical AI initiatives and can offer further discounts for projects aimed at making AI more equitable for under-represented peoples and mitigating dataset biases. 
<br>
__How does it work?__
- Step 1: Submit the form below with a description of the dataset you'd like to develop 
- Step 2: Our ML team will assess the request and follow up to understand the dataset size and specify the annotation task 
- Step 3: Scale AI will annotate the dataset


Scale for AI Research

Learn how to accelerate the development of AI applications with a secure annotation platform for document processing. Support a range of inputs from PDFs, Microsoft Word Documents, JPEGs and more.


Scale Document

PandaSet aims to promote and advance research and development in autonomous driving and machine learning. The first open-source AV dataset made available for both academic and commercial use, PandaSet combines Hesai’s best-in-class LiDAR sensors with Scale AI’s high-quality data annotation. It features:

- 48,000+ camera images
- 16,000+ LiDAR sweeps
- 100+ scenes
- 28 annotation classes
- 37 semantic segmentation labels
- Full sensor suite: 1x mechanical spinning LiDAR (Pandar64), 1x forward-facing LiDAR (PandarGT), 6x cameras, On-board GPS/IMU

Download PandaSet 

Getting started Scale AI? [Scale Docs](https://scale.com/docs/ "Scale Docs") is your go to resource page on how to get started with the data platform for AI and policy information. Sign-up to get the latest updates!

Scale Docs

Learn how to accelerate the development of AI applications with the standard-setting, advanced annotation platform for 3D sensor data. Capable of supporting large volumes of industry standard-setting training data leveraging machine learning-powered pre-labeling, tooling and quality assurance systems.

Scale 3D Datasheet

We [investigated the CoNLL-2003 dataset](https://scale.com/blog/if-youre-de-biasing-the-model-its-too-late)—a standard for building algorithms that recognize named entities in text—and found that the data is highly skewed toward male names. 
Download the Named Entity Recognition annotations for the original CoNLL-2003 dataset and our augmentation of it to reduce bias in your models. 

Balancing CoNLL 2003

Learn how to accelerate the development of AI applications with the most scalable annotation platform for video data. Capable of supporting annotations for a wide range of computer vision and classification applications, machine learning teams trust Scale Video to accelerate the development of high-quality models. 

Scale Video Datasheet

Learn how to accelerate the development of AI applications with the most comprehensive annotation platform for image data. Capable of supporting annotations for a wide range of computer vision, natural language and classification applications, machine learning teams trust Scale Image to accelerate and ensure the successful deployment of high-quality models.



Scale Image Datasheet

Learn how to accelerate the development of AI applications with a sophisticated annotation platform for text-based data. Support inputs from a diverse range of languages, machine learning teams and researchers and ensure successful deployment of unbiased NLP models.


Scale Text Datasheet

Products

Solutions

Resources