VIAM ML

AI & Machine learning pipelines for smart machines

A flexible, intuitive solution to manage any phase in the ML lifecycle, uniquely developed to deploy on machines running in the physical world.
An illustration showing the Viam platform being used to securely manage a fleet of smart machines
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.
A diagram explaining how Viam ML works
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 ant 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

A photo showing the Viam mobile app being used for fleet management
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.
WATCH THE VIDEO
An image of images being tagged in the Viam app
Blog - Nvidia and Viam
- raising the bar on
smart machine ML
Learn more about how Viam is partnering with leading hardware manufacturers like Nvidia to evangelize the power of ML on smart machines.
READ THE BLOG
Documentation -
ML model service
Check out our docs to label data, use your labeled data to train ML models, and deploy those models to your smart machines.
READ THE DOCS

Get started now

Viam is an open source platform, and developers can sign up for free and get started right away.
Looking for more of a guided tour? To see how Viam can take your enterprise’s fleet to the next level, fill out this form and one of our experts will reach out with a customized demo tailored to your hardware.