PhD Student Positions in Software Engineering and Business Process Management @ University of Tartu

Published on 15 May 2016

The Software Engineering Research Group at University of Tartu has four PhD openings in the fields of software engineering and business process management:

  1. Towards Greener Software with Software Analytics. Supervisor: Dietmar Pfahl
  2. A Data-Driven Recommender System for Software Test Automation. Supervisor: Dietmar Pfahl
  3. Structuring Business Process Models. Supervisor: Luciano García-Bañuelos
  4. Business Process Privacy Analysis. Supervisor: Marlon Dumas

The positions are suitable for students who have completed or are about to complete a Master degree in Computer Science, Software Engineering or related fields (the degree must be obtained by end of June 2016). All applications from eligible candidates are welcome but in case of high competition, preference will be given to applicants with high GPA in the Bachelors and Masters (e.g. GPA > 4 in the ECTS system); applicants with a high grade in their Masters thesis; and applicants who have demonstrated an interest in research during their studies.

All positions come with a scholarship of 1000 euros net per month for a full-time study commitment plus a possibility of earning additional teaching assistant income.

The Software Engineering Research Group conducts research and teaching in the fields of software engineering and information systems engineering, with a focus on process modeling and simulation, as well as software measurement, data mining and business intelligence. The group is one of the world-leading research concentrations in the field of process modeling and management and has strong expertise in data analytics for business process and software process management. Members of the group have earned 12 awards at leading international conferences in the past 10 years.

Applications from talented students from all backgrounds are encouraged. The application procedure is different for Estonian students and for international students:

  • International students should apply via the DreamApply system as instructed here. You should apply for the PhD in Computer Science @ University of Tartu. Please indicate the name of the supervisor (listed above).
  • Estonian students should apply via the SAIS system as instructed here. Please search for the PhD competition to which you wish to apply. The competitions can be searched by title.

In all cases, the application should include a so-called "research plan" (or "application letter"), addressing the following points:

  • The title of the project you wish to do (among the projects above). If you wish to apply for two of the above research projects, you need to submit two separate "research plans", one for each PhD project.
  • Your motivation and capability to carry out the PhD project proposed by the supervisor
  • Your personal ideas on how to successfully carry out and finalize the project
  • Your acquaintance with suitable methodology, the proper theoretical background of the applicant and familiarity with key literature
  • A short description of your previous academic activities and how they relates to the PhD project

The deadline for applications is 1 June 2016 for international applicants and 15 June for Estonian applicants.

Topic 1: Towards Greener Software with Software Analytics


During the last decade, green software development has become a hot topic and energy consumption has moved into the focus of software engineering. Traditionally, energy optimization research has focused at the hardware-level and at the system-level. Recent work indicates that there is ample opportunity to improve energy consumption at the software level. However, little is known about when software engineers are employing energy-efficient solutions in their applications and what solutions they employ for improving energy-efficiency.

This thesis project aims at extracting and analysing data from software repositories to learn from developers who successfully implement energy efficient software. Machine learning techniques will be used to extract and categorize qualified energy saving solutions for various types of projects and products. In addition, project results will be used to recommend energy saving solutions for defined contexts and to anticipate future trends in energy saving solutions across projects and application domains.

For further information, please contact Dietmar Pfahl (email:

Topic 2: A Data-Driven Recommender System for Software Test Automation


Test automation is a widely‐used approach to reduce the cost and increase the effectiveness of software testing. However, if test automation is not planned and conducted properly, it will not develop its full potential. Deciding what test activity to automate and what parts of a given System Under Test (SUT) should be tested automatically, at which level of intensity, and with what technique remains a challenge for test engineers and researchers alike.

This thesis project aims at developing a customizable test automation analysis framework which can be used as a recommender system with regards to what, how, and how much to automate along the software testing process in a given development context.

Recommendations will be based on data retrieved from software repositories. Machine learning techniques will be applied to generate recommendations. The final results of the project will include a prototypical tool and a comprehensive evaluation of the developed method and tool.

For further information, please contact Dietmar Pfahl (email:

Topic 3: Structuring Business Process Models

Business processes are often represented by means of conceptual models. Organizations value such models as important assets, because such models serve to understand the subtle and often complex sequence of tasks to accomplish the organization’s business goals. As the number of models maintained by an organization grows, the use of automated tools allowing stakeholders to consult, validate and redesign the business process models become unavoidable. In particular, it is desirable that such tools could help in ensuring the understandability of such process models. An important class of process models are structured in the sense that they can be decomposed in nested blocks, each block consisting of a single entry and a single exit point. There exists empirical evidence showing that structured models are easier to read and understand than unstructured ones, and that structured models are less error prone. Existing techniques for structuring process models can only handle acyclic models, including models with parallelism. Thus, the aim of this is to develop a generic and complete method for structuring process models in the presence of cycles with intertwined parallelism. The method should be parametrizable in the way that it should allow users to specify thresholds to trade off the structure and complexity (size of) the resulting model.

For further information, please contact Luciano García-Bañuelos (email:

Topic 4: Business Process Privacy Analysis

The aim of this doctoral project is to develop and evaluate techniques for privacy analysis of business processes. The main outcome will be a tool that takes as input process models with privacy metadata, and analyses these process models in order to: (i) detect unintentional disclosures of private information; and (ii) quantify the amount of private information leaked by the outputs of the business process. The tool will generate reports that explain to data owners the maximum extent of possible leakage of private data. The tool will also suggest possible counter-measures to reduce privacy leakages in a business process as well as the specific points in the business process where these counter-measures should be deployed.

This project is funded by DARPA's Brandeis program and is conducted in cooperation with Cybernetica.

For further information, please contact Marlon Dumas (email: