BytesAndBrains: A Shared Runtime for Machine Learning Outside the Data Center

Machine learning increasingly runs outside the data center, and the strategies the field has produced for that setting (federated learning, gossip learning, split learning, peer-to-peer inference) share little infrastructure: each ships with its own reference code, its own RPC stack, and its own training loop. BytesAndBrains is a runtime layer these strategies can share. It records a workload into an intermediate representation serialized as ONNX, partitions the recording across nodes, and manages the structured byte layer between them. The runtime is sans-IO: a state machine driven by the host program, with every distributed behavior testable in-process. The extension surface is composed of Components, typed plug-ins that satisfy Roles, runtime contracts the engine dispatches against. This paper describes the design, states its boundaries, and positions BytesAndBrains relative to Flower, FedML, NVIDIA FLARE, Ray, Substra, libp2p, TensorFlow Federated, Hivemind, and Spark.