Our mission is to build a trusted network of people and machines to annotate the world’s medical data.

Artificial intelligence is enabling extraordinary advances in healthcare. It promises to reduce costs and improve healthcare outcomes. But AI is only as accurate and effective as the data it is trained on. Nearly 30% of the world’s data is generated by the healthcare industry, but it is largely unstructured or poorly annotated, making it unusable.

Centaur Labs is advancing medical AI through trusted, accurate medical data labeling. Instead of relying on the opinion of a single medical professional, we harness the power of collective intelligence and performance incentives to achieve an unprecedented level of accuracy at scale. 

Why we do it

Patients and healthcare companies often trust a single physician’s advice, even though physicians disagree with each other at alarming rates. This disagreement, proven in numerous studies, is pervasive across clinical areas.

As AI plays an increasingly large role in healthcare, the consequences for bad quality inputs grow. How do you train an AI model when you lack confidence in the underlying data?

Misdiagnoses and errors in medical decisions—from AI and humans alike—cost lives. 

Collective Intalligence

Why Centaur

Centaur accurately labels medical data by intelligently combining opinions from experts with medical knowledge and specialized training.

Our approach is built on four pillars:

Brain with gears
Focus exclusively on medical data

We have a deep understanding of niche tasks, data formats, security constraints, and regulatory hurdles.

Brain with gears
On-demand global skilled network 

Tens of thousands of medical students and professionals compete against each other to label medical cases for our customers. On our labeling mobile app, DiagnosUs, we collect over a million opinions every day.

Multiple opinions are great
Performance-based compensation incentivizes the best work

We make the labeling experience competitive to constantly measure accuracy and incentivize effort from labelers.

Cloud with graphs
Collective intelligence drives improved accuracy

We intelligently aggregate opinions to create labels that beat the accuracy of individual board-certified doctors.

The stakes are high

40,000–80,000 patients die annually in the U.S. from diagnostic errors.

We know the stakes are high and that medical AI is hard, so we’re relentless about delivering the highest quality data labels.

We collaborate with and continue to invest in our network of experts. We listen to our customers’ challenges and customize solutions to improve their life-saving technology.

Our Team

We believe that a global network of people and machines is our best chance against the world's toughest medical problems

Want to join our team? We are hiring →

Our Fellows

We tap into an exclusive network of medical experts to ensure quality of your tasks

Our Advisors

Leading researchers are helping build Centaur Labs

Our Angel Investors

Alex Wang
Founder & CEO of Scale AI
Elliot Cohen
Co-founder & CPO of PillPack
Jake Klamka
Founder & CEO of Insight Data Science
John Capodilupo
Co-founder & CTO of Whoop
Kevin Carter
Former General Partner of SV Angel
Rick Morrison
Founder & CEO of Coordinate HQ
Rob May
Co-founder & CEO of Talla
Sunil Paul
Founder & Partner of Spring Ventures, Co-founder & CEO of SideCar
Tom Lee, MD
Founder of One Medical
Want to join our team? We are hiring →

Our Backers

Matrix PartnersAccel Ventures
Susa VenturesGFCOmega Venture PartnersY Combinator
MIT Sandbox Innovation Fund ProgramMIT Delta VhackvcGrant Park Ventures186 VenturesFAST