Build the map together
The best models are built with the institutions that care for patients. We’re looking for partners who want rigor, visibility, and real downstream impact.
Hospital & cancer center partners
We collaborate to build robust models and evaluate them honestly across institutions. The goal is shared learning — and tools that actually fit clinical workflows.
What we typically need
- De-identified whole-slide images (H&E; others optional).
- High-level clinical labels (diagnosis, stage, regimen, outcome).
- Optional: molecular data (RNA-seq / proteomics) for multimodal alignment.
- Clear governance (IRB / DUA), and mutually agreed data security posture.
What partners receive
- Recognition in publications and technical reports when appropriate.
- Early access to pilots; priority / partner pricing.
- Site-specific robustness analysis and model monitoring plans.
- Co-design of outputs and workflows with clinicians.
Pharma & biotech partners
We support target discovery and trial strategy through multimodal inference and mechanistic validation.
Biomarkers & stratification
Identify responder/non-responder signatures, stratify cohorts, and design inclusion criteria that increase signal.
Resistance mechanisms
Generate hypotheses that connect tissue phenotype to molecular pathways — then test the most promising ones.
Trial efficiency
Better stratification means fewer patients exposed to ineffective arms and a clearer path to efficacy signals.
Data governance & security
Trust is earned with process. We expect to formalize partnerships with appropriate legal and ethical frameworks.
Principles
- Minimum necessary data; de-identification where applicable.
- Clear data retention and deletion policies.
- Reproducibility: dataset versioning and audit trails.
Operational posture
- Secure storage, access controls, and logging.
- Separate research vs pilot environments.
- Shared documentation and review of data flows.