Machine learning that runs outside the data center.
A framework for machine learning that runs on phones, sensors, hospital networks, on-prem fleets, and peer-to-peer overlays.
Write your workload as a Module. The framework partitions it across nodes, binds your compute and transport, and ships the bytes.
What you can build on it
// worked example: a nervous system for machinesA fun example of a workload Bytes and Brains was envisioned around.
// in this example: system 1 acts at 120 Hz, system 2 reasons at 10 Hz, system 3 unifies the fleet from the cloud, and action latents gossip over the lan. every arrow is bytes the framework moves for you.
Built to spread across the edge
// three propertiesRuns off the grid
Compute lives where the data already is: phones, sensors, hospital networks, and on-prem fleets. No central cluster required.
Partitioned automatically
The framework splits your Module into per-node programs and schedules the pieces across a peer mesh for you.
Transport agnostic
Bind compute and transport independently. Swap between libp2p overlays, plain HTTP, or in-process calls without touching your model.
Roadmap
// what is nextTwo tracks: extensions to the framework itself, and protocols built on its primitives. The list keeps growing as the core stabilizes.
First public release: framework + sans-IO Engine + Wire + role traits + CpuBackend reference.
read the release notes →From the blog
// latest postAI, Velocity, and Trust
On AI velocity, losing the plot, and the trust levels I am using to earn back my own codebase.
read the post →Building in public.
The library is on crates.io. The design lives in the whitepaper. If you work on federated learning, gossip protocols, or peer-to-peer systems, reach out.
[email protected]