I AM AN
AI ENGINEER

Specializing in GenAI, NLP, and RAG systems. I design and build intelligent, scalable solutions, ranging from multi-agent extraction workflows to semantic search and real-time audio classification.

Experience

AI Engineer

TÜV SÜD Digital Service GmbH Munich, Germany

05/2024 – Present

Built award-winning RAG chatbots and knowledge management tools.
Orchestrated multi-agent document extraction pipelines, reducing bidding review times by 90%.
Accelerated service diagnostics 3x using a real-time voice-to-classification system.
Designed natural-language-to-SQL systems to increase retrieval quality.
Enabled agent communication and tool access via MCP and Agent2Agent integrations.
Automated user feedback and metrics reporting.

Machine Learning Engineer

Demos Plan GmbH Berlin, Germany

02/2023 – 04/2024

Prototype and benchmarked SOTA LLMs (including LLaMA) for product integrations.
Built RAG chatbots, semantic search, and recommendation systems for document analysis.
Led annotation and fine-tuned transformer models, boosting NER performance by 20%.

Data Scientist

rpc GmbH Munich, Germany

11/2021 – 11/2022

Created AI-driven customer personas and sentiment analysis systems using NLP.
Benchmarked ML/DL models to improve automotive sales forecasting by 10%.
Researched transformer explainability (BERT) to interpret model decisions.

Master's Degree

Technical University of Munich Munich, Germany

10/2020 – 10/2022

M.Sc. in Politics & Technology, with a specialization in NLP and Data Science.
Thesis Topic: Privacy Policy Question and Answering Systems with Multi-Task Learning.

Awards & Honors

TÜV SÜD

Innovation Awards (07/2025): 1st place for an AI-powered lift inspection solution and an enterprise chatbot platform.
Data Extraction Hackathon Winner (05/2025): Smart knowledge-graph pipeline for Bidding Management.

Portfolio

A selection of projects showcasing my creations in Generative AI, NLP, and Data Science.

xai streamlit transformers

XAI Privacy Intent Classifier

Explaining transformer model decisions when classifying user privacy intentions.

recommendation streamlit search engine

Movie Rec & Search

A semantic search and movie recommendation engine built with Streamlit and modern embeddings.

docker fastapi html/css/js

Privacy Policy Intent

A production-ready classifier dockerized with FastAPI and a clean front-end to identify privacy intent.

speech-to-text text-to-speech pytorch

AI Voice Chatbot App

A German-speaking voicebot combining audio classification, speech-to-text, and conversational AI.

gradio transformers scraping

TED Talks Summarizer

Scrapes TED transcripts dynamically and extracts concise, key insights using LLMs.

onnx quantization inference

Faster Model Inference

Quantized open-source models using ONNX formats to drastically speed up CPU inference times.

Skills & Tools

A curated selection of languages, frameworks, and technologies I work with to build intelligent solutions.

AI & Agents

LangGraph FastMCP Agent2Agent (A2A) AzureOpenAI OpenAI API LangChain Guardrails Vibe Coding (Claude, Copilot, Antigravity)

NLP & DL

Transformers SentenceTransformers PyTorch PEFT/QLoRA TRL Spacy / NLTK Speech-to-Text

Data & MLOps

Qdrant / FAISS FastAPI Docker Redis / Celery Dagster MLFlow / DVC Azure Cloud

Dev & Core

Python Scikit-learn Pandas / NumPy Streamlit / Gradio Web Crawling CI/CD & Git

Contact Me

Thank you for visiting my portfolio! If you have any questions or would like to connect, please don't hesitate to reach out.

I AM AN AI ENGINEER