Insider Brief
- Centaur Labs raised over $16M in a Series B funding round led by SignalFire, with participation from Matrix, Susa Ventures, Samsung Next, and Alumni Ventures, to enhance its health data labeling solutions for AI development.
- The company uses a network of over 50,000 expert annotators to label multimodal medical data, providing faster and more accurate annotations than traditional methods, supporting clients like Massachusetts General Hospital and Medtronic.
- Centaur Labs also launched an on-demand AI data labeling platform, enabling AI developers to receive expert-labeled data in under an hour, improving model monitoring and reducing risks in healthcare AI.
PRESS RELEASE — Centaur Labs, a leader in health data annotation, today announced the successful completion of a $16M+ oversubscribed Series B funding round led by SignalFire, with additional participation from Matrix, Susa Ventures and new investors Samsung Next and Alumni Ventures. In addition, Centaur Labs launched its on-demand AI data labeling product, which enables AI developers to collect labels from experts in less than one hour.
Funding will enhance health data labeling solution
While approximately 30% of the world’s data is generated by the healthcare industry, the majority of it is unstructured or poorly annotated. As AI and analytics play an increasingly large role in healthcare, the consequences for bad quality inputs or outputs grow. Many organizations struggle with the scalability, quality, and speed of data annotation necessary for model training and evaluation in healthcare, facing challenges such as low quality, management burdens, and difficulty sourcing skilled annotators. Generative AI models specifically can create an enormous volume of possible outputs, and it is up to humans to review those outputs for both factual accuracy (i.e., quality control) and preferred phrasing and framing of content (i.e., human preferences).
Centaur Labs helps health and science organizations accelerate AI development by using its network of over 50,000 experts to label multimodal data for model training, evaluation, monitoring, and feedback. Experts compete to be the best annotators on the company’s mobile app, DiagnosUs, ensuring only the most skilled annotators contribute to the data. As a result, Centaur Labs delivers greater accuracy and faster annotations relative to in-house teams or outsourced generalists — all on a HIPAA-compliant platform.
“In healthcare, where AI hallucinations can cost lives, ‘garbage in, garbage out’ data problems are unacceptable and models need to be ongoingly evaluated and monitored once they’re deployed,” said Erik Duhaime, CEO of Centaur Labs. “We see a future where AI, continuously refined by human expertise in real time, powers scientific breakthroughs and improves human health. This funding will allow us to build toward that future, further scaling operations and accelerating product development to ensure healthcare and AI companies can get medical data annotated with the quality, speed, and scale they need.”
“AI is radically transforming the U.S. healthcare industry, creating both immense opportunities and risk. Unlike other solutions out in the market today, Centaur Labs has found a way to deliver the scalability, affordability and accuracy needed to power AI models for medical devices, diagnosis tools, chatbots, drug discovery, and more. We’re thrilled to support Erik and the Centaur Labs team as they unlock the next wave of AI-powered health innovation,” said SignalFire Partner Yuanling Yuan.
Today, the company collaborates with AI leaders from startups to enterprises, with customers reporting up to 20X ROI from annotation speed improvement and the thousands of hours saved in data annotation tasks. Clients span the healthcare ecosystem and include industry leaders such as Massachusetts General Hospital, Memorial Sloan Kettering, Eight Sleep, Scibite from Elsevier, Activ Surgical, and Medtronic among others.
For example, Centaur Labs recently collaborated with Eight Sleep to advance the company’s snore detection algorithm by crowd annotating complex spectrograms and chest vibration waveforms to enable Eight Sleep’s Snore Detection model, boosting the model accuracy from 70% to 93%. John Maidens, Machine Learning Lead at Eight Sleep said, “The Centaur Labs platform helped us deliver this snoring model quickly. We went from conception to in front of customers in under a year, which is very quick for a biosignal AI.”
Launch of on-demand model monitoring, powered by experts at scale
Alongside the funding announcement, Centaur Labs will be alpha launching its on-demand data labeling platform. This new offering promises to deliver the fastest data labeling in the industry, with capabilities to serve urgent use cases in hours instead of weeks.
This launch is in response to the growing need for AI teams to integrate Centaur Labs ‘annotations into their data pipeline, and see a label turnaround on a daily or hourly basis for use cases like model monitoring. The on demand labeling platform will derisk healthcare AI in production and allow AI teams to identify opportunities to improve and catch high-priority failures before they have a negative impact. AI teams can load simple instructions and a single piece of data that — with the push of a button — is broadcast to Centaur’s entire network of annotators and receives 20–30 annotator reads in less than 30 minutes versus days.
About Centaur Labs
Centaur Labs annotates medical and scientific data at scale, helping AI developers to build, monitor, and validate their models. We offer a secure, HIPAA compliant annotation platform that supports annotation of multiple data types, as well as annotation services from our network of subject matter experts who provide millions of annotations weekly. Our platform leverages competition and performance-based incentives to ensure annotator skill and motivation, and intelligently combines the opinions of multiple people to generate the highest quality annotations. In addition, Centaur Labs partners closely with researchers at leading institutions all over the world to advance the field of AI in the medical and life sciences and has been published in publications such as Nature Digital Medicine.
