About me

I’m Alex Sheldrick, an AI/ML Engineer focused on Large Language Models and Retrieval-Augmented Generation systems. I hold an M.Sc. in Robotics, Cognition, and AI from TUM and an M.Sc. in Physics from TUB.

During my physics studies at TUB, I conducted research in experimental physics, contributing to publications in molecular spectroscopy. Later, at the Niessnerlab (TUM Visual Computing & AI Group), I focused on 3D computer vision research, including neural fields, neural radiance fields and gaussian splatting. During my work there I developed a novel probabilistic loss function for depth-supervised neural radiance field models, improving training efficiency and reconstruction quality.

At Aleph Alpha, I engineered LLM-based compliance tools, achieving significant performance improvements through systematic optimization. As a Data Scientist at ExoMatter, I built an LLM-based data ingestion pipeline and developed machine learning models for material science applications. While at appliedAI Initiative, I developed privacy-focused LLM middleware incorporating semantic caching and analytics.

Recent projects include developing a RAG-based historical storytelling system using ChromaDB and fine-tuned LLaMA models, and implementing IT helpdesk automation that reduced resolution times by 30%. I work primarily with Python, PyTorch, and cloud infrastructure (AWS, Docker, vector databases).

I’m fluent in German and English, with working knowledge of Italian, French, and Spanish. This has proven valuable in technical consulting across international teams.

Outside of technology, I’ve competed at the highest levels of E-Sports. In 2017, I founded and led a team to victory in the WoW-Arena World Championship, securing the $280,000 prize pool. I represented professional organizations including Tempo Storm during my competitive career.