T2R2
Enabling Data Scientists spending most of their time doing AI and Data Science
In AI and Data Analysis Workflows, Data Scientists spend most of their time struggling with computers, hardware and software, waiting for deployment issues to be fixed, dealing with annoying IT processes, talking to support, etc.
When they run complex big data processes they often end ud starting an analysis multiple times tweaking parameters until it eventually works.
T2R2 enables Data Scientists to focus on what they are good at.
Frictionless Deployment
Zero deployment effort, zero instantiation effort.
Once loaded in T2R2-One, any complex AI or Data Analysis module can be used in Research or in Development in compliance with all health regulations.
T2R2 enables focusing on Data Science without compromising anything.
Making Data Science Results easily Replicable and Reproducible
It might seem surprising as it’s all about computers and mathematics, but it is a well documented fact.
It’s a substantial part of the current general reproducibility crisis in Science.
AI and Data Analysis are cumbersome. Producing a Yes or No decision for a drug, a vaccine or a device for one patient or thousands implies running millions lines of code written in many different programming languages.
The smallest change in a single step can have a huge impact making a result not Reproducible.
T2R2 solves the problem at its core. It bridges the gaps between Research and Development.
Cloud-Native
T2R2 is built using last generation of Cloud-Native Software Architecture.
It ensures jobs distribution with the highest performance for the lowest cost and power consumption.
It scales up horizontally as well as vertically on demand.
It also ensures IPR of data as well as of calculation modules.
AI based
We are developing AI modules to optimize cloud resources usage for big data processing modules using historical data from previous runs.
Thus, the AI algorithms or data processing are optimized and workflows executions are continuously improved.
We are using advanced Probabilitic Graphical Models to produce similarity likelihood scores between results to ensure ultimate Reproducibility.
T2R2 by Design
Transparent, Traceable, Replicable, Reproducible (T2R2) by design
T2R2-One takes transparency to its deepest ground. Everything is tracked and nothing can be tampered.
The code execution becomes its own mathematical proof and is stored in a data store using Ledgers Technologies.
T2R2-One supports quality standards like GxP and 21 CFR Part 11 and it implements FAIR Principles as well.
Marketplace
T2R2 will provide a software market to enable anybody to deploy or use complex AI or Data Analysis modules.
It will become a hub to share AI and data analysis modules and to interact with their authors. It will enable a new dimension of data analysis, giving anybody access to the best scientific methods.