neumeric
neumeric is infrastructure for biological intelligence — the foundational layer between computational systems and biological cognition. it's an attempt to build the primitives that don't yet exist: interfaces, data pipelines, and architectures for systems where wetware and software co-evolve.
the gap
we have increasingly powerful AI and increasingly well-understood neuroscience. what we don't have is the infrastructure connecting them — APIs for the brain, pipelines for neural data, frameworks for building systems that operate at the biological-computational boundary. the tooling for this is fragmented, bespoke, and largely inaccessible outside of academic labs. neumeric is an attempt to change that.
what we're building
the current focus is neural interface data infrastructure: standardized pipelines for acquiring, processing, and acting on high-dimensional neural recordings. this includes signal normalization, real-time decoding, and closed-loop feedback systems that adapt to biological responses. the longer-term roadmap extends into model architectures specifically designed for the statistical properties of neural data.
why it matters
the next decade of bio-intelligence will be shaped by whoever builds the infrastructure layer first. neumeric is a bet that the right abstraction — clean, general, composable — doesn't yet exist and needs to be built from scratch rather than patched together.