Labelbox provides a collaborative training data platform that serves as a command center for data scientists and dispersed annotation teams. The platform integrates tools to annotate data, train models, conduct error analysis, refine annotations, augment or collect additional data, and repeat the loop to improve model performance. Rather than forcing companies to build homegrown tooling, Labelbox offers an end-to-end solution for the entire training data iteration cycle. The company is led by CEO Manu Sharma and co-founder and President Brian Rieger. Labelbox is used across industries including agriculture, insurance, healthcare, media and military intelligence, with customers such as ArcelorMittal, Chegg, Genentech and Warner Brothers. Financially, Labelbox raised $110M in a Series D and has to date raised $189m in total funding. The company intends to use the new capital to accelerate growth and production. Labelbox provides a web-based platform and API that lets data science teams and annotation teams manage labeling across images, text, documents, conversations, and video from a single dashboard. The product supports customizable tools (instances, custom attributes), access controls, labeler performance metrics, a catalog of labeling services, and feature/object analytics to improve models. Labelbox emphasizes pre-labeling with model predictions and active learning to reduce time and cost by automating high-confidence labels and surfacing low-performance assets for human review. The company was founded in 2018 by Manu Sharma and Brian Rieger and is headquartered in San Francisco. It serves roughly 150 customers and has just over 100 employees. Labelbox says the newly raised funds will be used to acquire new customers, expand its solutions, and grow its workforce globally, and its total capital raised is $79 million. Labelbox offers a platform that helps customers label and manage the data needed to train AI and computer vision models, including tooling to automate labeling for high-confidence predictions. The product supports workflows that let teams manually label only data below a prediction-confidence threshold. Proceeds from the recent financing will be used to double the size of its engineering and sales teams and accelerate the product roadmap. The company also plans to codify best practices and standard metrics for model performance by working with university and business partners. Labelbox emphasizes operational roles and standard data-exchange formats to speed AI deployment inside customers. The firm was founded in 2018, is based in San Francisco, and has about 30 employees; it has raised $39 million to date. Labelbox builds a collaborative training data platform that centralizes labeling, data management, and model-debugging workflows for computer vision ML teams. The company positions its product as the default software for machine learning teams to create and manage high-quality training data, likening its role to GitHub for software engineers. Customers include FLIR Systems, Lytx, Airbus, Genius Sports, KeepTruckin, and the company reports thousands of users worldwide after notable growth in 2018. Labelbox says the new funding will support additions such as automation, collaboration, and enterprise-grade features. The company plans to double its headcount in 2019, hiring across engineering, sales, marketing, and customer-success roles. Financially, Labelbox announced a $10 million Series A and reports $14 million in capital raised to date. Labelbox provides a software-as-a-service platform that lets human experts and crowdsourced labor annotate and manage training data for machine learning models. The product acts as an interface for teams to create many annotation types (e.g., gradations, matching systems) and applies data-science tools to reconcile reviewer discrepancies and surface edge cases. Customers include enterprises such as Lytx (DriveCam) and Conde Nast, and the company says it has over 5,000 customers; hundreds have tried the free tier. Pricing moves customers from a free tier to SaaS plans once they hit a usage threshold, with fees tied to the revenue the client’s AI generates. Long-term, Labelbox plans to manage fine-tuning data across an algorithm’s lifecycle so models can be continuously optimized. The team acknowledges a risk that many potential customers remain in R&D and that broader corporate adoption will determine the size and timing of market demand.

Total raised

$189M

Funding rounds

5

Latest round

Series D

Latest activity

Jan 2022