I am a physicist at KIT's Institute of Experimental Particle Physics. Among others I have contributed to the following open source projects:

• llama.cpp: a C/C++ program/library for large language model inference. Mostly contributions to the CUDA code with a focus on reducing hardware costs for consumers.
• kafe2: a Python framework for likelihood-based parameter estimation. Contributions were varied; the most significant one is the fast calculation of the correlated Gaussian likelihood via QR or Cholesky decomposition.
• fastNLO: a C++ library for the fast calculation of perturbative quantumchromodynamics cross sections at up to next-to-next-to-leading order via parton distribution function interpolation. I have contributed workflow and performance optimizations as part of my master's thesis.

If you are interested in contributing to llama.cpp but don't know where to start I will happily answer your questions via Mumble; just send me an email.

Emailing me about a temporary ban will result in a permanent one instead.