Why we invested in Centaur Labs, the modern AI data labeling solution for healthcare

The healthcare industry is poised for an AI transformation, but its progress is hindered by a lack of high-quality, labeled medical data. These datasets are critical for training AI models to make accurate medical predictions and decisions. The challenge lies in the complexity of medical data, which requires expert knowledge to label, must abide by the strict privacy regulations around healthcare data, and account for the variability in data across different sources. Without well-labeled data, AI models risk making inaccurate diagnoses or treatment recommendations, delaying innovation and care. 

That’s where Centaur Labs steps in, offering a platform that leverages a global network of medical professionals to annotate and validate medical data at scale. These labeled datasets are crucial for training AI models and enabling more accurate diagnostics, medical devices, and drug discovery processes. What sets Centaur Labs apart is its crowdsourcing model that combines medical expertise with collective intelligence to ensure data quality. This model is particularly relevant given the growing need for AI in healthcare, where precision is key and labeled data must meet rigorous standards.

Centaur Labs leverages a mobile app called DiagnosUs to gamify the data annotation process, engaging a global network of over 50,000 experts, including medical students and healthcare professionals. These experts compete to label data accurately, and the platform ensures quality by using hidden test cases to reward high-performing annotators. Centaur’s collective intelligence approach combines the best annotations while filtering out lower-quality ones. Their HIPAA and SOC 2 compliant platform processes millions of opinions weekly, scaling efficiently with smart tools and foundation models. This allows Centaur Labs to provide high-quality, on-demand annotations at competitive prices. Clients include Massachusetts General Hospital, Memorial Sloan Kettering, Eight Sleep, Scibite from Elsevier, Activ Surgical, and Medtronic among others.

What makes Centaur Labs even more compelling is the founding team’s unique blend of medical and technological expertise. Co-founder Erik Duhaime leads with a background in cognitive science and AI with a PhD from MIT in collective intelligence, focusing on building technology that can harness the power of human insight in tandem with machine learning. This interdisciplinary approach, combined with the team’s deep understanding of healthcare challenges, allows them to offer scalable, high-quality solutions that address the complexities of medical data labeling.

We’re pleased to join Centaur Labs $16M Series B funding round led by SignalFire, with additional participation from Matrix, Accel, Susa, Y Combinator, and Alumni Venture. With AI’s increasing role in improving patient outcomes, Centaur Labs fills a critical gap by ensuring that the AI models used in healthcare are trained on high-quality, reliable data, making the company a crucial player in the AI-driven future of medicine.


Rameen Rana is an investor at Samsung Next. Samsung Next’s investment strategy is limited to its own views and does not reflect the vision or strategy of any other Samsung business unit, including, but not limited to, Samsung Electronics.

Previous
Previous

Why we invested in Mawari, a DePIN network pioneering the next-generation of spatial computing

Next
Next

Why we invested in Vsim, a new physics simulation framework for robotics