Synthesize siloed patient data

Gain insights from real-time patient data that live across silos.
Win more data contracts from partners.

The federated data platform
for the life sciences

Silo is a platform that answers your questions
about clinical data that’s physically
on your partners’ databases.

Insights on live data

Train models across your connected datasets in real time.

Up To Date

The datasets are always up to date—no more waiting for data processing and upload.

Continuous

Continuously monitor trends and anomalies in the datasets.

Win more data contracts

Gain your partners’ trust via SiloML’s privacy-first approach to data management.

Local and Private

You can guarantee to your data partner that their data does not leave their premises.

Hassle-free

Your data partner does not need de-identify or upload their data.

How It Works

Data Partner

Register data sets

  • 1. Run server on machine with datasets
  • 2. Register server on the webapp
  • 3. Grant researchers access to datasets on the webapp
Run dataowner server

Researcher

Connect to datasets

  • 1. Request access to datasets on the webapp
  • 2. Run researcher server on machine they do data analysis on.
Run researcher server

Silo Platform

Orchestrate the learning

  • 1. We use Tornado to proxy private connections between the researcher and data owners.
  • 2. Under the hood, the federated learning occurs on the data owner and researcher servers via the PySyft framework.
Inspect Proxy code