👩💻 Currently:
Dipping my toes into Data Engineering with Databricks as a Working Student at Mercedes-Benz AG!!
🌱 Previously:
🔍 Built image classification pipelines during my Bachelors
📊 Explored clustering, feature engineering, and storytelling with data
🥣 Cooked up Agentic AI workflows at Talk2Scholars (✨ citation-aware RAG + scientific paper automation)
- 🕵️ RAG pipelines & LLM agents that actually do useful things
- 🛠️ Debugging — because nothing ever works on the first try
- 📈 Making dashboards and insights speak human
- ☁️ Pretending I know cloud stuff (Azure... I’m coming for you, eventually)
- Languages: Python (main), C++ (survival mode)
- Frameworks: PyTorch, LangChain, FastAPI, Streamlit, Scikit-learn, Keras, Tensorflow
- Tools: Git, VSCode, Jupyter, insomnia for APIs & occasionally sleep
- AI things: Transformers, CNNs, Embedding Models, Clustering, OCR, Retrieval systems
- Not AI things: SQLite, Pandas magic, beautiful plots, debugging print statements
🤖 AgentTeam (Master's Thesis) Building a multi-agent system that retrieves, transforms, validates and repairs ETL pipelines using LangGraph and structured AI workflows. Basically giving data ingestion pipelines a brain.
🚀 Talk2Scholars Citation-aware AI assistant using LLMs + metadata-aware retrieval. I don’t fine-tune LLMs (yet), but I make them talk smart.
🌤️ EventGo
Sniffs out sunny spots in Germany and finds nearby events using Open-Meteo & Ticketmaster. Kafka keeps it all in sync, MongoDB remembers, Docker ships.
🍽️ Chillar App
Mobile-first expense tracker with OCR for grocery receipts & eat-out logs. Minimalist UI with maximum sass.
- I write code like I cook: experimental but edible
- Fluent in Python and small talk in 3 languages (working on Italian 🇮🇹)
- Always learning, always slightly confused, always caffeinated ☕
"It’s not a bug, it’s an undocumented feature... unless it breaks everything."