NEXUS
The workbench for your own models.
Prepare data, then train, evaluate, deploy and monitor your AI models on a single platform designed for data and IT teams.
The platform
Four spaces, one tool.
Everyone works in their own space, on the same data and the same models. No more going back and forth between tools that don’t talk to each other.
Factory
Explore data, code in notebooks, train and compare models.
Launch pad
Put a model into service in a few clicks and expose it through a documented API.
Data governance
Know what data exists, who accesses it and which models use it.
Administration
Users, projects, compute resources and single sign-on (SSO) in one place.
Lifecycle
From dataset to live service.
Every step leaves a trace: you always know where a model comes from and how good it is.
Prepare
Connect sources, then clean, label and version your datasets.
Train
Run training jobs on CPU or GPU, in a notebook or as a scheduled pipeline.
Evaluate
Compare versions on the same criteria before you choose.
Deploy
Expose the model as an API, with its quotas and history.
Monitor
Watch quality and drift, then retrain when needed.
Compliance and explainability
Models you can explain.
A model that plays a part in a decision has to be justifiable. NEXUS documents each model throughout its life, with no extra work for your teams.
-
Automatic model card
Data used, parameters, evaluation results and versions, gathered with no manual input.
-
Bias and explanations
Spot the variables that weigh on decisions and any differences in how groups are treated.
-
Feedback from the field
User corrections go back to the model to improve the next versions.
Infrastructure
Where your data needs to stay.
NEXUS runs on a single cloud or several, in a hybrid set-up or on your own servers. Federated learning even lets you train a model on data that stays at each site.
- Private cloud
- Multi-cloud
- Hybrid
- On premises
- Sovereign cloud
- CPU and GPU usage tracked per project
- Scheduled processing and retraining
- Single sign-on with your company directory
- Federated learning across several sites
- Your models served through the same API as our models
Frequently asked questions
NEXUS in practice.
Who is NEXUS for?
Data, IT and business teams that want to build their own models and run them over time, without piecing together a dozen tools themselves.
What kinds of models can we build with it?
Forecasting, classification, anomaly detection and computer vision models, as well as language models fine-tuned on your data.
Can we keep our usual tools?
Yes. Notebooks and the common open-source libraries work in NEXUS, and your models remain exportable.
Do we need a data team already?
Not necessarily. Our engineers can run the first project with you, then hand over to your teams.
Demo, API access or workshop
Let’s talk about your project.
A specific question?
A NEO demo, access to the models API, a workshop on your data: tell us what you are looking for and an expert will get back to you.
sales@pertimm.com