PLATFORM
ALLONIA is a 100% AI platform open safe sovereign collaborative based on Federated Learning techniques. With ALLONIA platform, exploit the potential of Artificial Intelligence while maintaining the confidentiality of your sensitive information and ensuring the sovereignty of your data.
ALLONIA, the platform to accelerate and secure your Artificial Intelligence projects
ALLONIA is the collaborative platform that allows your technical and business teams to develop, deploy and maintain artificial intelligence models in a secure manner. Exploit the potential of Artificial Intelligence while maintaining the confidentiality of your sensitive information and ensuring the sovereignty of your data.
The Platform
OFFERS SEVERAL SPACES
- A “Factory” space to create and manage your objects: data, notebook, model, module, service
- A “Launch pad” space to manage the production of your jobs and services
- An administrative space to manage users and their roles
- A space to govern the data
BRINGS TO YOUR TEAMS
- Pre-trained and powerful AI models for different use cases
- Onboarding projects that make it easier to get started on the platform
- A collaborative space for the users of the platform
- Resource and storage management administrative capability
Build projects faster and put them into production in a few clicks. Using a sovereign infrastructure hosted in France, with complete security of access and exchanges. Find the data and models you want with our French and European sovereign software partners. Focus on creating value by collaborating with your team in an all-in-one environment. An AI platform made for your teams !
Benefits of the platform
An industrial platform designed to optimize the performance of your AI projects and facilitate their production.
Collaborative
We offer an optimized and collaborative user experience at each stage of development : exploration of AI models, creation and optimization of processing chains, management of working environments.
Federed
The ALLONIA platform was designed to provide controlled and tractable federated learning capabilities either between multiple projects within the same organization or between multiple companies to collaborate on model building without sharing the underlying data.
Simple
In a few clicks, easily integrate your data, create and train your models, deploy and monitor your projects. A user-centric design, always aiming to simplify your experience.
Open
The ALLONIA platform provides access to the main libraries of algorithmic models and federates a consortium of French and European sovereign editors among the best data providers, algorithmic editors and software publishers on the market.
Secure
Carry out your AI projects in a sovereign infrastructure, hosted in France, assuring complete security of your data. Access your space in a totally secure way and exchange with your teams, customers and partners.
Sovereign
ALLONIA guarantees you full control of the data thanks to an infrastructure entirely hosted in France (not subject to the Patriot Act or the Cloud Act) and offers you algorithmic processing, integrating open-source software (not subject to commercial licences).
Collaborative
We offer an optimized and collaborative user experience at each stage of development : exploration of AI models, creation and optimization of processing chains, management of working environments.
Federed
The ALEIA platform was designed to provide controlled and tractable federated learning capabilities either between multiple projects within the same organization or between multiple companies to collaborate on model building without sharing the underlying data.
Simple
In a few clicks, easily integrate your data, create and train your models, deploy and monitor your projects. A user-centric design, always aiming to simplify your experience.
Open
The ALEIA platform provides access to the main libraries of algorithmic models and federates a consortium of French and European sovereign editors among the best data providers, algorithmic editors and software publishers on the market.
Secure
Carry out your AI projects in a sovereign infrastructure, hosted in France, assuring complete security of your data. Access your space in a totally secure way and exchange with your teams, customers and partners.
Sovereign
ALEIA guarantees you full control of the data thanks to an infrastructure entirely hosted in France (not subject to the Patriot Act or the Cloud Act) and offers you algorithmic processing, integrating open-source software (not subject to commercial licences).
Workspace
Model management
Our model management allows you to track, compare, monitor, & verify model performance with ease throughout the model lifecycle.
You can monitor default generated metrics or manage custom ones, and check the drift of your models over time.
We provide you with all the necessary tools to ensure the performance, reliability, and especially the control of your models.
Performance monitoring
Comparison of models
Pipeline management
The management of your pipelines is simplified with our features.
Create reusable models from any part of your notebook or code. Assemble them and configure your pipeline with custom variables and scheduling options.
Monitor all runs, replay them, and be notified of their status at any time.
Pipeline creation
Follow-up of the executions
Model exhibition
Advanced service management makes it quick and easy to deploy and monitor any model or code.
With a few clicks, you can simply expose specific functions of your code or model through dedicated APIs, check the technical requirements and deploy your services on the fly.
A set of dedicated tools also allows you to monitor the performance and health of your services in real time, giving you total control.
Model exposure via API
Monitoring of active services
Compliance & explainability
We understand that compliance and explainability are essential for AI projects. This is why we have developed features to manage model compliance through a document automatically generated at the time of model creation and enriched throughout its lifecycle, centralizing auto-generated or customized metadata (validators, ethics rules, etc.).
Our tools allow you to use off-the-shelf explainer functions or to use your own, so that your models are no longer a black box.
We also provide tools that allow you to easily set up feedback loops to evaluate the functional value of your models based on your real usage data.
Enriched model documentation
Model bias monitoring
European Infrastructure
ON-PREMISE
If you want to deploy our platform on your own infrastructure ? Choose on site !