My team and I, with guidance from Microsoft Germany, developed a full-fledged chatbot web app for student services at TUM School of Business. The app assists with inquiries related to admissions, theses, and other student services.

Designed and developed a CLI that writes your git commits with AI.

Developed an open-source logging and analytics platform for OpenAI: Log your ChatGPT API requests, analyze costs, and improve your prompts.

Developed an escape maze game with path finding algorthim (implemented from scratch), character selection, levels, and gamification (achievements). Note: This project was developed into University GitLab ptivate repository, that's why there's only one commit on GitHub repository

As a student co-author on this research project, I played a key role in developing a robotic assistant aimed at enhancing ADHD therapy. I implemented the robot's hardware and software systems, including configuring a Raspberry Pi to integrate advanced language models like ChatGPT-4 Turbo and Claude-3 Opus. My work involved setting up the natural language processing capabilities and ensuring seamless interaction between the robot and the therapy environment. This project, validated through rigorous testing, has the potential to significantly improve personalized therapeutic practices for individuals with ADHD.

In 2023, I contributed to the Swift open-source project by adding a new filter method to C++ std::map and std::unordered_map, enabling these types to provide a Swift-friendly filter function with a similar interface to the Swift standard library's Dictionary.filter(_:). This enhancement improves interoperability between C++ and Swift, allowing developers to filter C++ maps directly into new instances of std::map while working within Swift. This contribution involved implementing the feature, addressing build issues, and ensuring the new functionality passed all tests, significantly enhancing the Swift ecosystem's C++ interop capabilities.