QA-Enhanced MLOps pipelines for robust and adaptive ML-intensive industrial systems

Sycodal is a Canadian company that empowers local small and medium-sized enterprises (SMEs) with cutting-edge digital solutions to overcome human resource shortages and adapt to market fluctuations. It is a member of the NVIDIA Inception program and uses NVIDIA’s technology to add machine learning (ML) and reinforcement learning (RL) capabilities to its automation systems. These systems, such as manufacturing robots and vision-based quality control, must meet high standards of service level agreements (SLAs). This project aims to develop machine learning model operationalization (MLOps) pipelines for ML-intensive industrial systems. These MLOps pipelines will include advanced quality assurance (QA) processes and methods to release ML solutions that can handle changes in industrial environments and naturally-occurring data shifts. As part of this project, two students will work on two different industrial ML applications, namely visual anomaly detection and self-learning robotic arms. Sycodal will adopt and integrate the results of this research project into its main ML system engineering workflow that is used to release vision-based systems and RL-based robotic arms for customers.

Faculty Supervisor:

Foutse Khomh

Student:

Partner:

Sycodal

Discipline:

Computer science

Sector:

Manufacturing; Professional, scientific and technical services

University:

Polytechnique Montréal

Program: