User:Pinkban/sandbox

= Zetane Systems Inc. (Company) = Zetane Systems is a privately-held Canadian company based in Montreal, Quebec that develops software tools for enterprise and industrial applications of machine learning. The company's software suite allows teams to debug, test, and track machine learning models to make them operationally ready for clients and stakeholders.

Early Years
Zetane Systems Inc. was founded and incorporated in 2016 by Dr. Patrick St-Amant, Guillaume Hervé, and Jonathan Magoon, who are the company's current Chief Technology Officer, Chief Executive Officer, and Chief Product Officer, respectively. Zetane Systems reported a revenue of $0 for the 2016 fiscal year, and has no published revenue for the years 2017 to 2020.

Growth
As of 2023, Zetane Systems is comprised of 17 employees, comprised of software developers, machine learning engineers, data scientists, and graphic/user interface designers.

For the 2022 fiscal year, Zetane Systems reported a revenue of $1.9 million, a $700,000 increase compared to their 2021 reported revenue of $1.2 million.

They have raised $1.3 million in funding over 9 rounds of non-equity assistance. Their most recent funding of $148,000 was received on June 10, 2022, from Hypercroissance Québec.

Products and Services
As of 2023, Zetane Systems offers 2 artificial intelligence-based SaaS platforms.

Insight Engine
Insight Engine is the first service deployed by Zetane System's. It is a tool that allows users to conduct deep model diagnostics on machine learning models. Insight Engine provides a workspace for understanding neural networks, and breaking up the layers of a machine learning model, with abilities to inspect internal tensors, network maps, and model outputs. Insight Engine is available as a Python API, or as a downloadable application.

Protector
Zetane System's main service is Protector, a tool for industries and enterprises to conduct automatic rigorous testing of their artificial intelligence and machine learning models to ensure robustness. The tests are comprised of weather, brightness, blur, transformation, obstruction, and colour tests that produce statistics on the inputted model's accuracy, precision, recall, and F1-score. Users can use the insights to analyze their machine learning data prior to official deployment. Protector is available as a Python API, or as a web-based platform.

Partners

 * PME MTL
 * Business Development Bank of Canada
 * Investissement Québec
 * Google
 * Amazon Web Services
 * District 3
 * Fasken
 * Creative Destruction Lab
 * The National Research Council of Canada
 * Économie et Innovation Québec
 * Mtl Inc.
 * Export Development Canada
 * Epic MegaGrants