Curriculum Vitae
Experience
AI Engineer (Freelance / TUM Social AI Club)
Architected and deployed an event-driven n8n workflow using the Google Drive API to automate audit verification. Integrated LLM pipelines to audit text against Spanish project-specific grant compliance requirements.
Working Student Data Scientist / Data Engineer
Designed LangChain RAG-LLM matching pipelines over OpenAlex publications. Managed local inference infrastructure using Ollama (Llama 3, Mixtral) on NVIDIA H100 GPU nodes. Migrated communication database from MySQL to PostgreSQL using pgloader under zero-downtime constraints.
Data Scientist / Machine Learning Engineer
Main developer for ICASSP 2025 challenge ensemble models combining EEG and Spotify features. Explored diffusion models for Spotify audio visualization and conducted a 50+ user study. Developed PyTorch pathology image artifact detection reaching 94% validation accuracy using OpenSlide.
Teaching Aide — Algorithms & Complexity
Provided tutoring and debugging mentorship to 200+ students across Java, Python, MATLAB, Haskell, and Prolog.
Intern Software Engineer
Ported legacy single-threaded TCP/IP communication modules from VB6 to modern C#/C++, resolving thread timeouts under heavy network load using asynchronous patterns.
Intern Data Engineer
Automated financial reporting extraction using Bash and PowerShell scripts, designing underlying SQL modeling to support PowerBI metrics for 200+ employees.
Data Engineer
Developed address normalization and scrubbing pipelines for logistics OCR address recognition systems.
Software Engineer
Worked in a Scrum team rewriting softstarter motor modeling tools utilizing MATLAB, C#, and VBA.
Education
M.Sc. Data Engineering and Analytics
Focused on Machine Learning, Big Data pipelines, GenAI RAG tools, and Multi-modal architectures.
B.Sc. Computer Science
Focused on Systems programming, Algorithms & Complexity, Logic, and Quantum Computing simulations.
Awards
ICASSP 2025 EEG-Music Challenge (2nd Place)
Ranked 2nd out of 50 international submissions. Built an ensemble model processing a 20GB dataset of EEG signals and Spotify metadata (published in IEEE proceedings).