Master Thesis

Master's Thesis Program 2027 in Data Science & AI (Also Available for Bachelor's Theses)

Working at Oktogrid™ Oktogrid™ is helping electricity providers bring the grid online for higher reliability and a safer energy transition. The future will be electric, and Oktogrid™ is providing the insights needed to keep the power flowing. We are a dedicated team working with industry heavyweights in optimizing the usage of the electrical grid and making sure that the electrification of our society remains sustainable. The Opportunity We run this Master Thesis program every year and for 2027 we are offering two thesis projects for Master's and Bachelor's students interested in Data Science, Machine Learning, and Artificial Intelligence! These opportunities are ideal for students eager to gain hands-on industry experience while completing their theses. Students will independently drive their projects under the supervision of a Technical Product Manager and work alongside SMEs, software developers and data scientists. In addition to fulfilling their academic requirements, the student’s project may be repackaged and integrated into one of our company’s products, directly contributing to its success. The Projects Project 1 – AI-Powered Release Notes Generation The scope of this project is to develop a locally running AI solution capable of automatically generating customer-facing release notes from screenshots of software user interfaces. Using Small Language Models (SLMs), image processing, and Optical Character Recognition (OCR), the solution should interpret screenshots and generate accurate descriptions of software updates. The generated content will then be organized into a predefined template and exported as a Microsoft Word (DOCX) document for human review. The project will explore the feasibility of automating technical documentation while maintaining consistency with Oktogrid's terminology and communication standards. Key technologies and concepts: - Small Language Models (SLMs) - Computer Vision and OCR - OpenCV and PaddleOCR - Automated Document Generation - Python and Local AI Inference Project 2 – Anomaly Detection in Transformer Measurement Data The scope of this project is to investigate Machine Learning techniques for detecting and classifying anomalies in electrical transformer measurement data. Oktogrid has collected over 5 million hours of measurement data for each of its sensor types, including electromagnetic field (EMF), microphone, vibration, and temperature. The project aims to explore whether Machine Learning algorithms can identify abnormal patterns more accurately than existing rule-based thresholds. The student will analyze historical sensor data, develop and evaluate anomaly detection algorithms, and investigate whether detected anomalies can be associated with specific transformer failure modes, including mechanical, electromagnetic, dielectric, and thermal faults. Key technologies and concepts: - Machine Learning and Artificial Intelligence - Anomaly Detection - Time-Series Analysis - Signal Processing - Classification Algorithms - Python and Predictive Maintenance Practical Details - Both projects are open to Master's and Bachelor's students. - The duration will be discussed with each student, but on average, we expect approximately 5 months of full-time effort for Master's theses and 3 months for Bachelor's theses. - The project can be conducted remotely (from home). The candidate is required to come to the office only once a week for a review meeting but is welcome to work from the office more frequently if desired. - A company laptop will be provided. How to apply Please apply with your resume in English and in PDF format by sending an email to ruben@oktogrid.io. No cover letter required.

No Matching Opportunities? Send Us Your Application!

If you don’t see a position that matches your skills or interests, we still want to hear from you. Feel free to submit an unsolicited application, and let us know how you can contribute to our team. We’re always on the lookout for passionate talent!

How to Apply

Please send a short introduction about yourself and your CV (in PDF format) to the following email address: jobs@oktogrid.io