Platform designed to manage end to end data challenge
Fully Managed Platform with Flexible Hosting Options
Our trustii.io platform is accessible from anywhere, no need to install, maintain or learn any software. We provide flexible data hosting locations and give option to select data scientists by university and region, so you can keep sovereignty on your data.
Automated Machine Learning based on Open Source
Trustii.io use best Open Source automated machine learning libraries, and use ensembling methods to create high quality models for demo and proof of concept. Trustii.io automated machine learning model is also used as baseline for competitions.
Machine Learning Challenges
While Automated machine learning is good for demo and proof of concept, Trustii.io can host competitions with students data scientists from top universities, where students compete on increasing the quality of the automated model and provide to organizations a model that meets their specific needs.
Future proof data scientists
The platform offers a real time leaderboard, where organizations can identify the best solutions and talented students, so they can contact them and hire them to continue the collaboration internally.
Trustii.io has integrated Forum where data scientists and organizations can communicate during the data challenge. Also Data scientists can collaborate with each other and submit their solution as a Team.
Integrated Jupypter Notebook (coming soon)
Optional Trustii.io hosted Jupyter Notebook for Data Scientists, for easy development, machine learning model code evaluation and data confidentiality
Model delivery and deployment
Once a model is available, you can deploy it at trustii.io cloud and create predictions on live data using our UI or API. Or you can download it and deploy it at your own cloud. trustii.io provides the model as a notebook that includes the code source.
Trustii developed own Machine Learning Ensembling methods that aims to increase performances and efficiency of machine learning model. Our Ensemble helps to reduce the variance of machine learning model, so the quality of predictions.
For us every model count in a challenge, and not only the the first three solutions.