My Projects

Designed and built a water-jet-propelled surfboard using primarily FDM 3D-printed parts (PLA). Integrated a waterproof electronics bay and a modular battery system; jet geometry was iterated in CAD for improved hydrodynamics.

Analyzed a real-world robotics demonstration dataset for connector insertion in cable assembly. Performed time-series preprocessing, value estimation (discounted returns), and visual analytics of human search strategies. Trained a Behavioural Cloning model (Keras) to imitate expert actions from state signals (pose, wrench, twist) for safer and faster policy initialization.

Built during the TUM.ai x Unite Hackathon, where our team achieved 1st place in Challenge 1 (Feature Normalization) and 1st place in Challenge 2 Level 1 (Core Demand Forecasting). The project focused on solving real-world procurement and recommendation problems using machine learning. We developed systems for scaling and normalizing product features across millions of items and predicting core demand patterns to improve recommendation quality.

Legacy AI is an AI-powered knowledge preservation platform that captures the knowledge your organization actually runs on. As a multi-tenant SaaS solution with role-based access, it combines AI voice interviews, automated knowledge extraction, and semantic search to transform tacit expertise into a searchable, living knowledge base — so organizations never lose critical know-how when key people leave.

As part of this research project at the University of Stuttgart, I worked on the development of an additively manufactured ceramic thruster for water-based satellite propulsion. The project focused on building a reliable manufacturing route — from material and process preparation to DLP printing, thermal post-processing, and experimental characterization.

Developing propulsion systems and lightweight components for next-generation space and aircraft using advanced materials.