What makes us unique
Octonion brings the power of Machine learning and Inference closer to where the data is generated
Learning and Inference directly in the microcontroller
High-performance models with a low footprint
Unsupervised machine learning:
Continuous learning and Dynamic clustering,
Prediction models with a Machine health score
Supervised machine learning:
Neural network and Classical ML algorithms inference,
Automated training pipeline selection
AI Studio with Zero coding approach
Training data in
time-series format
AI Studio with
zero coding
approach
High performance
models with a low
footprint
Learning and
inferencing directly
on microcontroller
How we solve challenges
We overcome challenges of different IoT domains using the best Artificial intelligence approaches
Keeping the same software stack
from prototyping to industrial
AI at the Edge
Secured
 connectivity
End-to-end
 software stack
“AI at the Edge” capabilities together with full IoT stack
Supervised and unsupervised AI learning approaches
Smooth migration from prototyping to industrial phase keeping the same
software stack
Zero code environment for AI models building
Rich visualization and insights on the AI model operation
Diversity of IoT connectivity options (BLE, LTE, LoRa)
Increased battery life due to decreased data traffic
Why AI at the Edge is important
Edge Artificial intelligence moves the AI closer to where it is really needed - in the device where action
happens locally
Optimized for ultra low bandwidth
AI at the Edge is necessary for ultra low bandwidth networks as raw data cannot be transmitted via these networks
Reduced power consumption
Reduce power consumption as only the relevant data is sent to the cloud, thus improving battery life
Better decision making
Enable smart, local and real time decision making
Less data storage
Reduce costs for Cloud data storage as less data is transmitted
More secured
Less vulnerable than streaming and storing data in the cloud
Reduced data consumption
Reduce costs for data communication since less data is transmitted
Challenges with AI in IoT
Customers are faced with different challenges when they want to apply AI in their IoT project
Get accurate data for
Machine learning phase
Monitor AI model
performance
Scale the AI model
Find the balance between
HW processing power and
inferencing efficiency
Get the right software
skills
Estimate properly the needed
development timing
Get in touch
Want to talk? Please feel free to send us a note