Viam’s customizable platform goes beyond one-size-fits-all to tailor AI & ML modeling to individual machines, unlocking new ways to improve their performance.
A cloud-based AI & ML system specialized for individual machines
1
Let machines capture your data and configure when and what to send from machines to the cloud.
2
Label data and train ML models for each machine’s unique conditions.
3
Deploy ML models onto individual machines and write code that allows your machines to optimize their performance in response to current conditions.
BENEFITS
AI & ML tailored to each machine using any data or model you want
Leverage ML infrastructure dedicated to enhancing smart machine capabilities
Bridge the gap between machine data capture and ML tools created with cloud data centers in mind. Turn information collected from machines into intelligence that can optimize business outcomes.
Selectively capture any relevant image or sensor data for training
Collect the data that is most relevant to your needs on any camera or sensor, without wasting resources or including irrelevant information.
Flexibly take advantage of ML on your terms
Whether you’re just starting with smart machines and ML or want to make what you already have more efficient, with Viam ML you can capture, label, and train on any data you want, and deploy any model on any machine.
FEATURES
Continuously make machines smarter based on the data they collect
Simple model labeling and filtering
Easily organize the data that you want to train models on and filter it within the Viam app to focus on the information that matters most.
Rapid ML testing and iteration
After you train your models, you can test them against your datasets or deploy them for testing on a machine. Once you get a sense of how they perform, you can iterate on them by adding changes to your dataset and retraining your models.
Train or deploy existing models
Use Viam ML’s built-in modeling capabilities or upload your existing models for deployment across your fleet. Seamlessly build on top of your pre-existing architecture without altering your core setup.
Selectively capture any relevant image or sensor data for training
Collect the data that is most relevant to your needs, from any camera or sensor, without wasting resources or including irrelevant information.
Lightweight computer vision based on any ML model
Enable your machine to better understand the world around it through computer vision. Send data-intensive object recognition workloads for processing to the cloud to boost machine performance further.
Modular flexibility
Pull models easily and run them on machines quickly—it’s easy to try and swap out ML models for your machines. If what you need is not there, it is easy to add models to the registry and apply them to your fleet.
Use cases
Predictive maintenance
Smart home automation
Autonomous inspection
Environment monitoring
Security through facial recognition
Creative LLM integration
The power of a unified platform
Viam ML customers reduce the need to patch together various solutions and gain new efficiencies as they apply machine learning models to their fleets in conjunction with other solutions in the Viam Platform.
Viam Data
Readymade pipelines to capture, sync, store, and visualize machine data which then be used to train ML models to deploy to your machines.
Viam Fleet
Use insights gleaned from observing how your ML models perform in the field to update code, push it your fleet, and optimize performance.
Viam Registry
No need to reinvent the wheel - leverage the open source community for ML models or create your own to deploy to all your machines.
See ML in action
Viam’s application and ML model service make training and deploying machine learning models to your smart machines much faster and easier. Check out this brief video, where our developer advocate Arielle Mella demonstrates how to deploy ML on a smart machine through Viam in less than 5 minutes.
Further resources
Demo - How to selectively capture data for machine learning using Viam
Don’t waste valuable resources on data that won’t enhance your machine’s performance - capture only the data you need and use it to train your ML models.
.png)
