Hello!

Data for frontier AI.
Progress for people.

We’re building high-quality datasets for AI training and evaluation, with a focus on complex domains where context and specialized knowledge matter.

We’re interested in problems that call for reasoning across sources, interpreting multimodal evidence, and evaluating decisions in context. Our goal is to connect real-world data with the expertise needed to make it useful.

• Real-world datasets with clear provenance
• Domain expertise to guide curation and evaluation
• Training and evaluation tasks grounded in evidence

Our team.

We’re a team of former machine learning engineers, founders, and biological scientists. We obsess over data and its impact on safe, reliable AI.

An initial focus: medical data.

We already partner with contract research organizations (CROs) and biotech companies working to accelerate vaccine development.

Our medical data focus also includes oncology: clinical records, imaging, pathology, and longitudinal outcomes, connected across the patient journey.

For sensitive data, our approach starts with clear permissions, respect for privacy, and a voice for the people behind the records.

We’re looking to collaborate with frontier AI labs, research teams, domain experts, and data partners — in medicine and beyond.