Edge AI as it
should be...
Every Individual Unit
Do you know how every unit you have out there is doing? The only place to learn this is in operations by not only doing inference but also doing training at the edge. Experience genuine individual condition monitoring.
Adaptive Automation
Programming handles known circumstances but how do you handle unknown faults and changing conditions? By automatically learning and adapting to the actual conditions of course. It’s time to switch from reactive to proactive.
AI on Everything
Struggling to get AI to run onboard your products at the edge? Ekkono is designed for deployment – super-tiny footprint, hardware and sensor agnostic, and model life-cycle management (MLCM). Run one solution on all your products.
Solutions
Condition Monitoring
Train a virtual sensor of a critical health indicator based on the individual unit’s normal behavior. Use predictions to detect before it exceeds a critical threshold and anomaly detection to catch deviations.
Condition-Based Maintenance
Turn Condition Monitoring into actionable insights like remaining time to service or remaining useful life (RUL) on wear parts. Using real-time data at the edge enables individual servicing and a transition from reactive to proactive support.
Performance Optimization
Run real-time simulations of alternative settings based on normal behavior and sensitivity analysis to automatically ensure that every unit always runs at its best for where, how, and what it’s being used.
Sustainable Products
Ensuring that every unit runs at its best and only when needed, which can only be done in operations, is the next big sustainability leap – a competitive advantage as your customers scope their environmental footprint.
Embedded Edge AI
With a super-tiny footprint, C and C++ libraries that run in any environment, on any hardware, with any sensor data, no dependencies, and hotswappable ML models, Ekkono is purpose-built to be embedded at the edge.
Software Development Kit (SDK)
Supporting a workflow from the application engineer, to the data scientist, to the embedded software engineer, this is the toolkit for model generation, optimization, evaluation and implementation, that runs where you want to run it.
Ekkono Synthesis
Going into production requires model life-cycle management (MLCM). Synthesis aggregates models to learn collectively from what has been learnt individually through model correlation, outlier detection, federated learning, and more.
TensorFlow Lite/Micro is designed for deep learning tasks like image classification on edge devices. However, the vast majority of edge ML use cases focus on sensor data—such as pressure, temperature, or power consumption—rather than images. These scenarios require functionality like incremental learning, real-time data preprocessing and anomaly detection, which TensorFlow Lite/Micro lacks.
"Though open source solutions are free, they lack dedicated commercial support"
Ekkono is purpose-built for these edge ML tasks, providing all the essential capabilities in a comprehensive toolbox. If you need deep learning for images, TensorFlow Lite/Micro is a great fit. For real-world edge applications, Ekkono delivers the tools needed to make devices smarter and more adaptive.
Trusted by
