During my time at university I took part in several hackathons:
MorningByte (hackaTUM Challenge Winner 2021)
MorningByte is an automated morning briefing system that detects the most relevant anomalies and trends from end-of-day sales data, generates narrative explanations with priority scores, and produces short videos for managers to consume bite-sized insights; the pipeline includes data analysis, text-to-speech, visualizations with Matplotlib, and video assembly with MoviePy, all orchestrated in Python and surfaced via a mobile-friendly app. Source / Repo
nourishLens (GDSC Ideathon Winner 2024)
nourishLens was developed during the GDSC Solution Challenge ideathon to address nutrition inequality by inferring dietary patterns from grocery receipts and offering cost-effective, healthier alternatives for low-income households, eliminating the need for manual meal entry and framing malnutrition as a problem of dietary quality and inequality. Live demo
nourishLens won first place at both the local Munich GDSC Ideathon and the Europe-wide competition.
ThinkTankGPT
ThinkTankGPT dynamically creates virtual expert personas to debate on any topic. Users can input a topic of their choice, and the system dynamically generates relevant experts who autonomously engage in a debate. Additionally, users have the option to join the discussion, directing it into their desired course or challenging the AIs viewpoints. The application was implemented in Streamlit using LangChain and was inspired by the MAD Framework to surface diverse viewpoints.