Tokyo, Japan (July 16th, 2025), EdgeCortix® Inc., a leading fabless semiconductor company specializing in energy-efficient Artificial Intelligence (AI) processing at the edge, today announced that its advanced MERA™ compiler and software framework is powering Renesas’ newly unveiled Renesas Unified Heterogeneous Model Integration (RUHMI) framework. RUHMI is an innovative toolchain designed to enable seamless and efficient deployment of the latest AI models across Renesas microprocessors (MPU) and microcontrollers (MCU), in a machine learning framework agnostic manner.
The RUHMI Framework offers developers a robust, modular solution for AI inference tasks across a wide range of Renesas MPU and MCU products, including the recently introduced RA8P1 MCUs. Powered by the EdgeCortix MERA compiler and software framework, RUHMI enables swift integration of neural network models with native support for popular formats such as TensorFlow Lite, Pytorch and ONNX. This integration significantly simplifies the deployment of customized BYOM (bring your own model) and ready-to-use applications across embedded environments.
“This initiative builds on our longstanding partnership with Renesas,” said Dr. Sakyasingha Dasgupta, Founder & CEO of EdgeCortix. “It validates our software-first strategy and underscores the transformative role of MERA compiler in simplifying AI deployments across heterogeneous processor environments. By abstracting hardware complexities, MERA ensures rapid, efficient, and scalable AI capabilities, spanning from MCUs and MPUs to neural processing units (NPUs) and advanced AI accelerators. With MERA powering RUHMI, I believe Renesas customers will increasingly benefit from its seamless support for heterogeneous compilation across both microprocessors and microcontrollers.”
Featuring integrated developer tools and comprehensive support for industry-standard model formats, RUHMI empowers embedded system designers with the flexibility and speed required to integrate intelligent functionalities into real-world industrial, automotive, and IoT products.
For more information about the RUHMI Framework and developer resources, please visit: