From 21c3d9c72f5e452fef723b74e2d9eec4605a4514 Mon Sep 17 00:00:00 2001 From: idky137 Date: Tue, 14 Apr 2026 13:26:37 +0100 Subject: [PATCH] updated readme --- README.md | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) diff --git a/README.md b/README.md index 1c9f328..a6459b5 100644 --- a/README.md +++ b/README.md @@ -6,22 +6,30 @@ [![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE-APACHE) [![MSRV](https://img.shields.io/badge/MSRV-1.81-blue.svg)](#) -> **Portable, contract-enforcing computation graphs for AI-enabled embedded -> systems — from bare-metal microcontrollers to multi-threaded servers, in -> safe Rust.** +> **A real-time runtime for continuous AI systems operating on live data +> streams — portable, contract-enforcing computation graphs from bare-metal +> microcontrollers to multi-threaded servers, in safe Rust.** > **Alpha release.** Core contracts, code generator, and integration-tested > runtimes are functional. APIs may change before v0.1.0. +> See [umbriel-systems.com/limen](https://umbriel-systems.com/limen) for +> project overview and target domains. + --- ## The Problem -Building AI-enabled robotics and edge computing systems today means -rebuilding the same software pipeline for every hardware target. A -perception-inference-actuation pipeline written for a Cortex-M4 cannot run -on a Raspberry Pi or a server without substantial rework — despite -implementing identical logic. +AI systems are increasingly deployed in environments where data is +continuous, latency constraints are strict, and resources are limited. +Robotics platforms, industrial control systems, and edge sensor networks +require real-time processing of live data — not batch jobs or API calls. + +Existing infrastructure — from PyTorch to ONNX Runtime — is designed for +batch or request-response workloads. These systems introduce unpredictable +latency, dynamic memory behaviour, and limited control over execution. And +building for a Cortex-M4 means rebuilding everything when you move to a +Raspberry Pi or a server — despite implementing identical logic. No existing framework bridges this gap: