Directory of Experts
Back to search results

Research project title

MLOps - Support the development and operations of machine learning applications

Education level



Director: Marios-Eleftherios Fokaefs

End of display

May 31, 2025

Areas of expertise

Artificial intelligence

Software and development

Computer systems software

Software engineering

Computer systems organization

Unit(s) and department(s)

Department of Computer Engineering and Software Engineering


The candidate needs to provide evidence of competence in at least 3 of the following subjects:

1. DevOps

2. Machine Learning

3. Cloud computing

4. Software Configuration Management (Version Control)

5. Software evolution and maintenance

6. Data science


To apply, the candidate should contact the professor with the following information:

1. The word "MLOps" in the subject of the email.

2. Proof of competence in the above subjects (courses, projects, papers).

3. CV

4. Transcripts

5. A critical review of this article (strengths, weakness, open questions) 1-2 pages.

Detailed description

Software configuration management is a formal process that guides the development and evolution of the system (IEEE, 2012). Among others, the process includes tasks such as version control, change management, "builds", processes, and faults. SCM is part of the general software quality assurance (SQA) process. DevOps specifies several tasks, sub-processes and tools to implement and facilitate the SCM for continuous integration, continuous deployment and continuous delivery. Currently, the SCM and SQA of software and AI models are separate, which does not allow the full potential of AI applications to be fully exploited. For this reason, our goal is to integrate and combine SCM and SQA with software and AI aspects against quality indexes and version control. When talking about business, it’s important to look at software from two aspects: the technical aspect and the economic aspect. The economic aspect of the software depends on its quality, which represents a view of its technical aspect. The quality of the software affects the perception of users and customers and, by extension, the economic power of the software product in a competitive market. Regarding AI applications, they consist of several models and several software components, sometimes not owned by a single entity. To assess the economic value of applications, we must combine AI quality and software quality and translate the quality into economic terms. By having the relationship between quality and economics of the software, you can help make decisions about the evolution of applications or the composition of AI models to develop applications. The specific objectives of the DevOps sub-project include:     1. Develop a platform for software configuration management (SCM) for applications and AI models.         a. Extend existing version control systems, such as GitHub, to include functionality relevant to AI.         b. Develop visualizations and other formats focusing for model information.         c. Connect the technical details of models and applications.     2. Develop decision support systems (DSS) to guide the development, evolution and composition of AI models.         a. Create hybrid models to capture the economic and qualitative aspects of AI models and applications.         b. Develop multi-objective optimization methods to support decisions while optimizing quality and economy at the same time.

Financing possibility

The project is part of a Mitacs project with the Thalès Group and the AI@Centech accelerator. The candidate must complete one or two units (4 to 12 months) of internship with the partner. Applicants are encouraged to apply for competitions for scholarships as additional funding.

Marios-Eleftherios Fokaefs

Marios-Eleftherios Fokaefs

Assistant Professor

Main profile