Industry partner
Netlight / UNHCR

Problem: 
The united nations refugee agency protects the rights of millions of refugees worldwide and support emergency protection and assistance for people forced to flee. The current communication ways between the UNHCR and people of concern (POC) is labour intensive, lacks a clear visibility of status of individual cases and does not provide self-service possibilities to POCs. As a consequence, POCs lack trust in the process, feel increased insecurity and face issues in communicating with their points of contact at the UNHCR. 

Solution: 
The students worked on mobiles applications for iOS and Android to support POCs in approaching the UNHCR and communicating about their cases. The applications support over 20 languages and can even be operated by illiterate POCs. The application not only enables POCs to make their case in a self-service fashion, but also increases transparency in the process and for the case workers at UNHCR. These student's work successfully demonstrated the feasibility and value of mobile lightweight and accessible applications to the UNHCRs mission. 

GitHub:
UNHCR Mobile App

Industry partner
Siemens

Problem
Nowadays every modern industrial software development project employs automatic analysis of security properties. While this enables the identification of security flaws, their communication and response are still an open issue. The challenge consists of designing a system, that supports developers independent of their security expertise to quickly identify and implement solutions for detected security flaws. 

Solution
Students developed a comprehensive security findings recommender system. The LLM-based system automatically analyses security findings and guides developers with a concise as well as a detailed explanation in the resolution of the security flaw. 

GitHub
Security Findings Recommender System

Industry partner
Itestra

Problem
Business systems are on of the daily drivers of enterprises. As such a bad system performance not only impacts employee satisfaction, but also business goals. A common cause of bad system performance is ill-conceived implementations, such as blocking API calls within loops. While it is easy to avoid them during implementation, once implemented their effective and efficient detection requires tremendous effort. 

Solution
Students implemented an extensible plug-in for Intellij to detect performance anti-patterns in java code through dynamic analysis and abstract syntax trees. The plugin focused on the performance anti-pattern of many database requests inside a loop, but can be extended to the other performance anti-patterns. 

GitHub
Anti-Pattern Analysis

Industry partner
Siemens

Problem
Security compliance heavily relies on scarce security expertise and by that hinders fast development. Compliance utilities are not streamlined with modern software development hindering integration of compliance into software engineering workflows. The challenge is characterised by bringing security compliance closer to engineers and facilitate assessments by non-security experts. 

Solution
Students designed and implemented a website providing an overview of the software artefacts necessary for compliance to IEC 62443-4-1. This enables engineers to easily access compliance utilities and get an understand of compliance as well as to preliminary assess their compliance posture without consulting a security expert. 

GitHub:
Security Compliance Assessment

Industry partner
Netlight / Karevo

Problem
Sorting potatoes is a tedious manual task, especially for small and medium-sized farms, as automated systems are costly and tailored to large-scale operation. Karevo addresses this by developing an affordable AI-driven potato3 sorting system. One important step aspect of such a system is the clear communication of configuration options and sorting decisions, accommodating farmer with non-technical background. 

Solution
The students developed a web framework enabling the configuration of the AI driven decision making, giving farmers insights into the configuration options and respective decisions. 

GitHub
Potato Evaluation Framework

Industry partner
Itestra

Problem
Staffing personnel to projects is a complex task involving various aspects such as prevalent workloads, available skill set, or employee preferences. Hence especially fast growing companies face trouble scaling their staffing approaches, leading to intransparent and inconsistent decisions, risking inefficiencies and dissatisfaction inside the company. 

Solution
Students engineered a platform to facilitate the staffing process. The platform introduces visibility into the individual employee's workloads, skill set and preferences as well as wishes for future projects. The platform not only enables an efficient and effective staffing process, but also to allows employees to improve their skills sustainably in new projects. 

GitHub
Project Staffing Plattform