Product Overview

The saying “If all you have is a hammer, everything looks like a nail” summarizes a cognitive bias we have to use tools that are most familiar to us, even if they are the wrong tools for the job.

Hence, one should naturally ask the question of what is the best tool or the best Hub of tools for doing Data Science. Ideally, it should be flexible enough to encompass a variety of tools, while unifying them onto a single platform to boost productivity.

A few of the typical roles targeted in Tru AI are:

  • Not a programmer, but still eager to do some Data Science experiments? GjirafaTech Tru AI provides a guided, easy-to-use set of tools so that you can quickly build a prototype.

  • Do you like Python? GjirafaTech Tru AI includes JupyterLab and smoothly integrates with the platform itself.

  • Interested in implementing the end-to-end ML pipelines with special emphasis on feature engineering and one click-deployment? GjirafaTech Tru AI is the answer since the special focus is given to these aspects.

  • A typical programmer who wants to use API-ready retrained SOTA models, aka AI Engines, such as Recommendation, Lookalike, Propensity Engine.

GjirafaTech Tru AI is a scalable, secure, easy-to-use platform. It provides a single shared workspace for team collaborations, with numerous productivity-enhancing features:

The Feature Engineering - functionality offers data scientists the option to prepare datasets that are compatible and fulfill machine learning algorithm requirements through an easy-to-use UI, improving the performance of machine learning models.

Data Sources - connect to different data sources, including databases such as MySQL, MS SQL, and Data Management Platforms such as GjirafaTech Bisko.

Artifacts - users can upload their datasets as CSV, XLS, TXT. Artifacts are also outputs of Datasets, Feature Engineering.

Models - users can train the ready-to-use datasets or artifacts with built-in algorithms, such as linear regression, SVC, RecSys, Apriori, and Collaborative Filtering, and many more.

Deployments - a trained ML model can only make predictions when it is deployed, therefore a deployment component exists on Tru AI, so its end users are able to consume the trained model.

Sharing of data/processes/models - when appropriate, with fine-grained control of users and groups

Schedules and queues for long-running processes on powerful server hardware

With ♥ by the Tru AI team © GjirafaTech 2022