Andreas H.Private profile: Upwork shows its earnings only to signed-in members, so we check it every few weeks, not daily. MRR here is a rough estimate (Upwork's six-month earnings ÷ months active). Last read Nov 5, 2026.

LLM & Retrieval Architect | RAG · Context Engineering · Vector Search | PhD I design LLM and retrieval systems that work in production. I was building context-engineering architectures before the term existed. I start by understanding your problem. Sometimes the answer is agentic RAG. Sometimes a simpler pipeline is better and cheaper. What I have done Designed the algorithms for two successful production services at a major affiliate-marketing company (topic modeling, semantic search, content attribution) Developed, in my own research, an exact vector search algorithm with perfect recall (R@k = 1, for any k. QPS=700 in the tail; measured to 500M vectors, on my laptop) Built a video segmentation and search pipeline for content creators (transcripts, scene and silence detection, semantic and keyword search) Detected rare events in large text collections for International Security Studies research Searched for new biomedical compounds for a biotech startup Consulting I also give go/no-go reviews. These help clients decide whether to start a project, or how to bring a stuck project to completion. Key technologies LLMs (ChatGPT, Claude), RAG and agentic RAG, vector search (FAISS, HNSW), embeddings, topic modeling (BERTopic, HDBSCAN, UMAP), spaCy, Hugging Face transformers, PyTorch, Python Experience Scientific Advisor, Geodesix by Impact Senior Data Scientist, Impact Senior Data Scientist, Trimes Principal Scientist, Nokia Research Center Education PhD in Physics & Natural Sciences (Rijksuniversiteit Groningen), Marie Curie Fellow. 10+ patents, 30+ peer-reviewed articles.