Edge Impulse and Ceva (CEVA) Collab to Boost Edge AI Applications for Faster and Easier Development
PRISM NextGen Tech and AI Index company Ceva, Inc. (CEVA) is the leading provider of silicon and software IP for Smart Edge devices, and Edge Impulse, a top platform for developing, deploying, and scaling edge machine learning models. Today Ceva announced a collaboration that is expected to boost Ceva’s NeuPro-Nano NPUs by integrating support for the Edge Impulse Platform. As a result, AI developers can now develop, train, and deploy advanced embedded ML applications for Ceva NPUs with cycle-accurate performance, even before physical hardware is available. This unique collaboration allows developers to benchmark and optimize the performance of their edge AI models in the Edge Impulse Platform, significantly reducing the time to market for NeuPro-Nano-based silicon aimed at AIoT devices.
“We are at the dawn of the Edge AI era, with growing demand for on-device AI across all industries and markets,” said Chad Lucien, Vice President and General Manager of the Sensors and Audio Business Unit at Ceva. “To meet the need for low-power, small form-factor, and powerful AIoT devices, our customers need a solution to efficiently incorporate embedded ML applications into their NeuPro-Nano-based products. The combination of Edge Impulse Platform and NeuPro-Nano offers a powerful solution, giving AI developers a seamless and user-friendly experience to rapidly design, develop, and deploy silicon-optimized AI applications with market-leading performance.”
The Edge Impulse Platform enables optimized AI performance for NeuPro-Nano NPUs with minimal to no coding required. Developers can upload or create models on the Edge Impulse Platform for faster development cycles and streamlined deployment on NeuPro-Nano. With the NeuPro-Nano architecture supporting AI inference, feature extraction, and DSP workloads—key for audio, voice, and vision applications—the integration with Edge Impulse ensures developers get accurate cycle count emulation for full end-to-end AI applications.
