This Master project is offered by the Configuration Management department of ASML and was developed together with EAISI (Eindhoven Artificial Intelligence Systems Institute, https://eaisi.tue.nl/).
Department: Configuration Management
If the journey regarding overlapping changes (multiple changes impacting the same item within the same timeframe) has taught is one thing, it is that the analysis of it was very challenging because the data was not interconnected. Process and product data were loosely coupled at best. That made analysis complex and a highly manual activity and process mining difficult. In many cases historic data could only indicate a potential overlap but we could not be 100% conclusive about it. In context of a highly complex product with many parallel running changes, the impact of delays are amplified because for every issue we face we need to do a lot of manual work to analyze data and be successful at finding the cause and mitigating it. In other words, the analysis is not easy to repeat and therefore we lose a lot of valuable time and do not have a clear operational view on the actual risks we need to deal with.”
This assignment is a study into the usefulness of event knowledge graphs for Configuration Management at ASML and identifying what use cases it can support. The objectives are to
- Create a lab environment where events can be recorded in a graph,
- Identify use cases that otherwise cannot be supported based on business input, and
- validate its value.
Event knowledge graphs are a way to model events as objects having relations to other objects like change objects or even product data like parts and datasets. This connects the process (through events) with the resulting product documentation explicitly. By bringing events as objects to the Configuration Management Baseline, analysis and process mining becomes easier as the data across tools will then be connected to all the events that have or will take place. In the Baseline we can even model the future events that have not occurred yet. Which we can potentially use to predict possible conflicts like overlapping changes from occurring in the future.
Profile of candidate
- Preferred educational background: Data Science and Artificial Intelligence
- Educational level: master
- Type internship: Graduation (master thesis)
- Required experience and knowledge of hard skills: experience with graph DB technologies, experience with Neo4J is preferred, basic understanding of event knowledge graphs (as taught in 2AMI20 Advanced Process Mining)
- Required personal skills: able to communicate and collaborate with both IT and business stakeholders, pro-active and self-organizing
This graduation project covers, both, the preparation phase of 3 months and the graduation project of 6 months (DS&AI and CSE since AY2021-2022).
- Preferred start date: 01-Nov-22 (preparation project)
- Days per week: During preparation: Nov -Feb: 28 hours per week, During graduation project: Feb – Aug: 40 hours per week
- Duration of internship: 9 to 10 months
Dr. Dirk Fahland (email@example.com)