Eleni Stroulia: Taking Computing from Research to Real Life 

The world of computing is no longer confined to screens, code, and technical systems. Its influence now reaches into healthcare, education, business, homes, and the everyday lives of people who depend on technology in ways they may not always see. The most meaningful advances often happen when technical knowledge meets a real human need, whether that means helping someone live more independently, giving students better tools to learn, or finding smarter ways to manage complex systems. 

At the intersection of these possibilities stands Dr. Eleni Stroulia, Professor in the Department of Computing Science at the University of Alberta. Her career has been shaped by a practical question: how can computing, AI, and machine learning be used to address problems that matter beyond the laboratory? Her research has consistently followed that direction, focusing on industry-driven challenges and developing technologies that can improve or automate tasks in meaningful settings. 

From 2011 to 2016, Eleni held the NSERC/AITF Industrial Research Chair on Service Systems Management with IBM, extending her work into the relationship between research and industry. Healthcare has become another important part of that journey through the Smart Condo project, where she investigates how technology can help people living with chronic conditions remain independent for longer, while also helping health-science students learn to provide better care. The Smart Condo team received the University of Alberta Teaching Unit Award in 2011. 

Her leadership has extended beyond individual research projects. Eleni has held leadership roles in the GRAND and AGE-WELL Networks of Centres of Excellence, received a McCalla Professorship in 2018, and was recognized with a Killam Award for Excellence in Mentoring in 2019. Having supervised more than 60 graduate students and postdoctoral fellows, she has watched many of them move into successful academic and industrial careers. 

Since 2020, she has served as Director of the University of Alberta’s AI4Society Signature Area, helping advance interdisciplinary work around artificial intelligence and its impact on society. Since 2021, she has also served as Acting Vice Dean of the Faculty of Science. Across research, teaching, mentorship, and leadership, her work reflects a belief that the value of technology is ultimately measured by what it makes possible for people. 

Where Mathematics Meets Language 

Eleni’s path into computer science began with two interests that might initially seem quite different: mathematics and languages. Mathematics appealed to her because of its ability to produce answers that could be confidently proven correct. Language offered something different, the ability to communicate, explain ideas, and persuade when there was no single clear answer. 

That combination eventually led her to computer science. At the time, she saw it as a young field with enormous potential to transform almost every aspect of life. She later realized that it also offered a way to bring both of her interests together. 

Algorithm design, for example, requires an understanding of computational complexity, with well-defined measures that allow different approaches to be compared accurately. Complex architecture presents a different kind of challenge. Design decisions often require people to argue for the relative merits of different options, explaining how each one responds to the requirements at hand. Those decisions are more open-ended and cannot always be measured so neatly. 

That balance between precision and interpretation has remained an important part of her academic journey. 

From Metacognition to Software Engineering 

Eleni’s research career did not follow a straight line. Her PhD thesis focused on metacognition in learning and explored a simple but powerful idea: an agent that keeps a record of its reasoning while solving a problem can learn from an unsuccessful outcome. 

When the result differs from what was intended, the agent can reflect on that divergence and identify the steps that contributed to the failure. Eleni describes this ability as distinctly human. People regularly look back at what they have done, consider what went wrong, and use that understanding to approach similar situations differently in the future. 

After joining the University of Alberta as an Assistant Professor, she brought this way of thinking into software engineering. Her work adopted the idea of “blame assignment for redesign” within evolutionary software development. By extracting different models of software structure and behavior, her team developed tools to help developers migrate software to newer technologies and redesign systems to improve efficiency and maintainability. 

Modernizing Legacy Systems Without Losing Their Core 

One of the first major results of Eleni’s work in model-driven software engineering was a methodology for moving legacy user interfaces to modern technology stacks. 

Legacy systems can remain highly valuable even when the technologies behind them have become outdated. They continue to provide the functions organizations depend on, but maintaining or extending them can become increasingly difficult as their original developers retire and older programming languages lose support. Meanwhile, users continue to expect the usability and accessibility offered by modern technologies. 

Eleni’s research addressed this gap by developing a systematic methodology and tools for creating modern user interfaces and middleware capable of translating user actions into calls to the underlying legacy system. 

The approach relied on a model of the legacy interface behaviour, built by bringing together and abstracting examples of how the system was used. Studying that model made it possible to redesign the interface around newer technologies and improve the user experience without having to compromise the core legacy system. 

Reading Software as It Evolves 

Another significant contribution of Eleni’s software engineering research involved differencing design-level software models extracted through static analysis of object-oriented software. 

Object-oriented languages give developers several ways to reuse software, supporting goals such as performance, modularity, and understandability. Eleni’s team developed tools capable of identifying opportunities for redesign, explaining the trade-offs between different alternatives, and automatically restructuring code according to the goals of system owners. 

The work reflects a broader interest in understanding not simply what software does, but how it changes and how those changes can be guided more effectively. 

When AI Becomes Part of the Engineering Team 

AI has introduced a remarkably flexible new tool into software development. Eleni has consistently built and evaluated her research methodologies through tools, and she now sees AI agents as capable of carrying out a wide range of software-development tasks in considerably less time than a human. 

Their capabilities, however, come with important limitations. AI-generated code can be unnatural, less elegant than code produced by an experienced developer, and difficult to understand or debug. More concerning is that AI agents may not always respect the constraints within which they are expected to operate, creating potential risks for software systems and their users. 

This is changing the role of people in software engineering. Rather than simply performing the task themselves, humans increasingly need to evaluate the behavior and outputs of AI agents, understand what they have produced, analyze it, and correct it when necessary. 

That shift also changes what future computer science students need to learn. Eleni believes they must be prepared to use AI effectively while also approaching its outputs critically. Alongside education, she sees a need for legal governance frameworks and testing protocols capable of determining whether AI-developed software is trustworthy and ready for deployment. 

There is another requirement that she considers both the most important and the most difficult: having mechanisms that allow experts to switch off systems when they misbehave. 

Software is now woven into physical infrastructure and everyday life so deeply that shutting down major systems can be difficult to imagine. Yet as AI agents become increasingly capable of learning and acting independently, another possibility becomes easier to imagine. In some situations, living without an AI-powered system may be less disruptive than accepting the risks of allowing it to continue operating. 

The Real-World Questions Behind Better Technology 

Eleni’s work has always moved between developing software-engineering methods and actually building software systems. Each side informs the other. The technical challenges that emerge while building systems often lead to new research questions, creating a cycle between practice and research. 

Her interdisciplinary collaborations have taken that approach into very different settings. With colleagues in Business and Education, she built software for social-science research in virtual worlds. Work with Computing Science, Occupational Therapy, and Industrial Design led to ambient intelligence systems that could unobtrusively observe and analyze the daily activities of vulnerable people. In collaboration with Computing Science, Medicine, and Nursing, she worked on smartphone and wearable software designed to extract indicators of cognitive and mental health. 

Those experiences have also shaped two major research projects she is pursuing today: digital building twins that use sensor data streams to support more efficient building management, and virtual-reality exergames designed to motivate personalized exercise protocols for different forms of frailty. 

Eleni sees interdisciplinary collaboration as one of the strongest ways to translate rapidly evolving computing knowledge into real-world applications. It requires computing scientists to understand the practical requirements of the environments where their systems will be used and to clearly define the boundaries of a system’s behavior and the quality of its outputs. 

That demand for explainability remains one of AI and machine learning’s difficult problems. Models can achieve impressive accuracy, yet their complexity and stochastic nature can make their decisions difficult to explain or reproduce. This has driven Eleni’s recent work on modular software pipelines for transparent and replicable experimentation. These systems document the complete data-processing and machine-learning process, support AutoML for algorithm selection and hyperparameter optimization, run specified experiments, and archive the resulting products for detailed inspection. Such platforms make it easier to replicate, compare, and extend research. 

Putting People Ahead of the Technology 

Eleni’s approach to responsible, human-centred AI has also been shaped by her leadership at the University of Alberta. The university’s Signature Areas initiative was created to bring together interdisciplinary research groups working in areas of institutional strength. As Director of AI4Society, her role has been to encourage interdisciplinary research and teaching, develop training programs, provide institutional leadership, and represent the university in initiatives involving public, private, and international partners. 

One part of that work involved creating seed funding for interdisciplinary teams beginning new collaborations. The resulting projects brought together areas such as arts and machine learning, mathematics and business, and engineering and medicine. These combinations often produced insights that would have been difficult to predict in advance about how AI should serve people. 

Bias is one such concern. Historical biases within data, combined with the objectives that machine-learning models are trained to optimize, can create several interacting forms of bias. Identifying and reducing those biases must come before deployment. 

The questions now extend further. Researchers also need to consider how AI systems handle values and moral dilemmas, including situations where a model should refuse to answer because the request could cause harm. The growing amount of personal data available to these systems also raises questions about ownership and ethical use. Academic research operates under established ethics protocols, but comparable safeguards do not necessarily follow when research products move into commercial deployment. 

A Personal Understanding of Belonging  

Being a woman in a technical field has given Eleni a personal understanding of impostor syndrome. Failure is part of every researcher’s experience, but there were times when she questioned whether failure meant she did not belong in the field. 

A supportive family and a strong education helped her develop the self-knowledge and resilience to move past those doubts. Growing up in Greece also gave her a deep respect for inquiry and a desire to pursue knowledge. Those values continued to shape her academic path and her sense of belonging, even when she did not often see people like herself in her professional environment. 

That experience influences how she approaches new working relationships. Rather than relying on assumptions based on nationality, skin color, or gender, she tries to understand people as individuals, including their character, background, values, and goals. 

Good Leadership Leaves Room for Good People 

Eleni sees a strong team as essential to the success of everyone involved, from junior developers to senior academics. She has been fortunate to work with capable and trustworthy people and to delegate responsibilities with confidence that they will be carried out. 

Delegation, however, does not remove the need for coordination. Her approach is to distinguish between tasks that are fun, urgent, and important, then prioritize them while maintaining a healthy mix. That balance helps create space for meaningful work alongside the demands that come with academic leadership. 

Failure is part of that process too. When something goes wrong, she believes the important step is to recognize the failure, learn from it, and try not to repeat it. 

A Legacy Measured in People 

Eleni finds the idea of legacy somewhat overwhelming. Yet she believes the most meaningful legacy anyone can create is found in the opinions of the people they have encountered, particularly students in an academic career. 

Her advice to the next generation is grounded in lessons she has gathered along the way. Have a plan, she suggests, but remain open to opportunities, because a detour can sometimes lead to the same destination. Give every task the full benefit of your skills and look for something interesting even in ordinary activities. 

For Eleni, that search for meaningful work is closely connected to joy. Finding something interesting, applying oneself to it, and becoming successful at it can be a steady path toward a fulfilling career.