Labelbox Inc.
Labelbox builds reinforcement-learning environments for frontier AI labs and managed-agent infrastructure for enterprises. Its products help teams create training data, evaluate models and agents, and deploy agents that learn from real work.
At a Glance
- Frontier AI labs and hyperscalers
- Enterprise AI and machine-learning teams
- Software and technology companies
- Healthcare and life sciences
- +4 more
AI Tools by Labelbox Inc.
(2)Labelbox
Managed AI Agents and Training Environments
Alignerr
AI Training Data Labeling Platform
Discussions
No discussions yet
Be the first to start a discussion about Labelbox Inc.
Latest News
Labelbox raises $100M+ Series E at a $1.25B pre-money valuation
Launch of Recursion Managed Agents for enterprise work
Introduction of Recursion, an enterprise agent reinforcement-learning platform
Welcoming Upcraft to the Labelbox team through an acquisition
Products & Services
AI data and model-development platform for collecting, cataloging, annotating and evaluating image, video, audio, text, document, geospatial, conversational and multimodal data; includes workflows, model-assisted labeling, human evaluation and APIs/SDKs.
Reinforcement-learning environment and evaluation offering for frontier labs, with simulated work environments, trajectories, expert-written graders, deterministic verifiers and preference/taste signals.
Enterprise reinforcement-learning platform for developing, evaluating and deploying specialist agents that improve from real enterprise execution.
Managed multi-agent fleet for engineering, research, reliability, security, product, finance, sales and other recurring work; agents run in sandboxes and connected business applications, use specialists and graders, and retain memory.
Market Position
Labelbox positions itself as an end-to-end AI data, evaluation and reinforcement-learning infrastructure provider rather than a standalone annotation vendor. It competes with data-labeling and AI-data platforms such as Scale AI, Appen, CloudFactory, Sama and Amazon SageMaker Ground Truth, and with model-evaluation/agent platforms as it expands into Horizon and Recursion; its differentiation is the combination of expert data, RL environments, verifiable grading, enterprise integrations and agents that learn from execution.
Leadership
Founders
Manu Sharma
Aerospace engineer educated at Embry-Riddle and Stanford; previously co-founded Ardulab/Infinity Aerospace, worked at DroneDeploy, and was a product manager at Planet Labs. He is Labelbox's co-founder and CEO.
Brian Rieger
Technical co-founder with an aerospace/software-engineering background; he studied with Manu Sharma at Embry-Riddle and became Labelbox's early co-founder and COO/president.
Daniel (Dan) Rasmuson
Technical co-founder with an aerospace/software-engineering background; he previously worked with Manu Sharma on the DroneDeploy app marketplace and was an early Labelbox technical leader/CTO.
Executive Team
Manu Sharma
Co-founder & Chief Executive Officer
Aerospace engineer and former DroneDeploy and Planet Labs product/engineering leader; co-founded Labelbox in 2018.
Brian Rieger
President & Co-founder
Technical co-founder with an aerospace and software-engineering background; previously Labelbox COO.
Board of Directors
Founding Story
The founders had built machine-learning systems and experienced first-hand that existing tools were not designed for producing and managing training data. They initially conducted extensive customer discovery and built the data-labeling tools they wished they had; the product quickly attracted users and revenue, prompting them to raise venture funding and scale. The company later built Recursion for its own internal agent workflows before offering it to enterprises.
Business Model
Revenue Model
Enterprise software and managed services sold through the Labelbox/Horizon platform, API/SDK and data/evaluation workflows, plus Recursion Managed Agents and expert-data services. Public materials emphasize enterprise contracts and a self-serve/API product surface, but do not publish a standard price card.
Target Markets
- Frontier AI labs and hyperscalers
- Enterprise AI and machine-learning teams
- Software and technology companies
- Healthcare and life sciences
- Financial services and insurance
- Media and entertainment
- Training and post-training of foundation models and frontier AI systems
- Reinforcement learning and RL from human or verifiable feedback
- Multimodal, conversational and generative-AI evaluation
- Computer-use and tool-use agent training and evaluation
- Software engineering, code review, debugging, dependency upgrades and flaky-test remediation
- ML research, experiment management and regression investigation
- Genentech
- Warner Bros.
- Black & Decker
- Stryker