I'm doing my MSc in Data Science at the University of Augsburg in Bavaria, Germany. I build AI applications with large language models: agents that use tools, RAG systems over private documents, and machine learning models served as tested APIs. Before my Master's I worked for about three years with data, analytics and automation in Germany, Canada and India.
- Agents and multi-agent systems with LangGraph and LangChain
- Retrieval augmented generation with local open models (Ollama, Llama 3.2)
- Turning models into services: FastAPI, Docker, GitHub Actions
| Project | What it does | Built with |
|---|---|---|
| react-data-agent | Answers questions about sales data by calling analysis tools. Runs as one ReAct agent or as a supervisor with two specialist agents | LangGraph, LangChain, Ollama, pytest, Docker, GitHub Actions CI/CD |
| rag-document-qa | Answers questions from your own text and PDF files, with source and page citations, fully offline | LangChain, FAISS, sentence-transformers, Llama 3.2 |
| automl-feature-pipeline | Customer churn model selected with AutoML (ROC AUC about 0.85) and served as a REST API | FLAML, LightGBM, FastAPI, Pydantic, Docker Compose |
- Deutsche Bank stock price prediction with an LSTM network
- Credit score classification comparing supervised learning models
- Bank customer churn prediction with a neural network
- Sentiment analysis on tweets with NLTK and XGBoost
- Restaurant clustering and review sentiment with KMeans
- Job market dashboard in Power BI
Python, SQL, LangChain, LangGraph, FAISS, Ollama, FastAPI, Pydantic, Docker, GitHub Actions, pytest, scikit-learn, LightGBM, XGBoost, PyTorch, pandas, Power BI, AWS
Looking for a working student position in AI or data science in Bavaria.