Lunch seminar: Towards Continuous Delivery of ML/DL Systems

Abstract: Artificial intelligence (AI) is becoming increasingly popular within companies across domains due to the success of machine learning (ML) and deep learning (DL) technologies. For companies across domains, the end-to-end process of developing, deploying and successfully adopting ML/DL models in operation introduces several challenges related to the design, the evolution and the scaling of these models. For example, access and availability of data is often challenging and activities such as data collection, data cleaning, data pre-processing, data storing, as well as uploading, training, deployment, and placement of model(s) in operation is complex. This is true particularly for large, complex, highly regulated and safety-critical embedded systems in areas such as e.g., telecom, automotive, defence, security, healthcare etc.  As part of the research, we conduct studies covering the end-to-end process of ML/DL model development, as well as in-depth studies focusing on the specifics of the deployment phase and how to operationalize these models in large-scale embedded systems, and how to continuously deliver and evolve Machine Learning Systems (MLOps) to improve the key quality attributes of the larger system of which these are part.

Bio:  Meenu Mary John is a PhD candidate at Malmö University. Academic background in Computer Science and professional experience in the software industry. She conducts research in Data- and Al-driven Development. The intension is to enable and advance not only expert data scientists, but also to encourage less-experienced data scientists, software engineers and even non-experts to effectively approach the design and evolution of ML/DL models in today’s embedded systems industry where the shortage of highly skilled data scientists is a fact. The research is conducted as a multi-case study research within the context of the Software Center.

Presentation from the meeting (pdf):
Towards Continuous Delivery of ML/DL Systems, by Meenu Mary John