The Peltarion Platform – the operational AI platform for producing real-world applications at scale and at speed

Peltarion’s operational AI platform is a single, unified platform, capable of executing a broad range of use cases and methodologies. Providing tight integration with other organization systems to enable cohesive, robust, auditable and scalable AI solutions.

The Peltarion Platform

02/ Platform benefits

AI development cycles for prototyping through to staging and production made exponentially faster.
Platform abstracts the complexity of the development cycle but allows for the application of powerful deep learning techniques. Intuitive GUI (with zero coding) allows more user types to work effectively with AI/deep learning.
Use a variety of data and AI/deep learning techniques. Infrastructure agnostic: operate in different cloud environments across different hardware types. Can use pre-built models or build from scratch.
A sophisticated end-to-end platform, with all the tools. Open APIs allow access to platform processes to enable automation into enterprise workflows and pipelines.
Audit past activity, map the data path through models to results and monitor real-time view of projects: status, costs, outcomes and progress to date.
Share, learn and build on the work of others; enable cross-functional teams to improve the end solution through collaboration and parallel experimentation.
Enable more users and less specialized teams with access to the platform to integrate AI into your organization. Ability to have fewer and less expensive resources working in shorter time spans on projects.

03/ Peltarion Platform: Views

Projects view allows you to create a new project, select, edit or delete an existing project.
Datasets view allows you to import, preprocess, augment and manipulate data to run experiments. The view gives a full overview of all the datasets available to the user, and gives access to information about the available datasets.
The modeling view area allows you to build an AI deep learning model in one place. You can start from scratch or select from a series of preloaded snippets, build a model yourself or work with a colleague on the same one. There is no coding required, just drag and drop to build your model.
In the evaluation view, you can see in real time how the AI model is performing as it’s learning from the data. You can see the results of multiple experiments at the same time, so it’s easy to compare different models. The view lets you analyze in real time if the experiment is going in the right direction or not.
When a model has been built and trained, it is time to deploy. Currently, you can deploy as API in the deployment view.

AI is driven off the data, and the right data. Sitting with the enterprise is generating the data. 

Access data from multiple sources.  Prepare data for modeling: visually explore, assemble, find, access, clean, combine and structure data.

Designing, training and optimizing the models to find the best fit for business objectives. Easily run multiple experiments in parallel.

Analyze all tasks, activities and outcomes across all areas from data preparation, modeling and deployment.

Best model promoted to production. The model is then accessed by the organization via an API.

Set of collaboration functions which allow team members to work and interact across an AI project.

Capability of overlaying the platform for organization management, reporting and KPIs, allocating and configuring resources, projects, roles and payments.

Full security capabilities to secure data and manage access to the cloud-based operational AI platform.

Cloud-based application extends to multiple providers and geographies.

Model accessed via API. Feedback loop as use of model in live production will lead to another cycle as the original model gets updated and improved.

04/ More detail

Get started in no time
Cloud based with no hardware or software installation. Easy and secure ingestion/integration of data source. Overhead free.
Develop AI services fast
Efficient workflows, assisted experimentation and an easy-to-use interface: drag-and-drop editors, pre-made AI models, parallel experiment execution and automated model serving.
Powerful experiment management
Compare and evaluate model performance. Track and manage experiment versions, status and outcomes. Transparent audit path reduces effort and errors.
Advanced and intuitive modeling
Building and testing models is quick and flexible. Bug-free modeling reduces the chance of errors. The latest techniques and research are all embedded into the platform.
Rich data capability
Get the most from your data with data management and augmentation. Bring images, audio and video into AI models. Ability to experiment with external sources and formats. Organizations own their data in a secure environment.
Enable team collaboration
GUI supports team collaboration. Tutorials, demos, best practice and case examples will guide users. Share across the project or organization.
Hardened for the enterprise
The platform is reliable, secure and easy to maintain. Workflow makes activity auditible and easy to reproduce for upgrades or new projects.
Run AI projects at scale
REST APIs allow for integration into reference or existing data pipelines. Platform capabilities can be integrated at scale into existing enterprise workflows.

05/ Use cases

06/ Get in touch

Contact us today.

We’d love to hear from you:

Your message has been sent.

Try the platform