Moritz Boos#
Moritz Boos
Head of Data Science at Fab AI · London, UK
I build and evaluate LLM and machine-learning systems.
At Fab AI, I lead individual data science projects evaluating language models for education in low- and middle-income countries, working on benchmark design, rubric methodology, LLM-as-judge systems, and the engineering needed to run them in production.
My background is in Bayesian statistics, causal inference, deep learning, and computational cognitive neuroscience. I use that background to treat LLM evaluation as a measurement problem: designing evaluations that are reliable, interpretable, and statistically sound.
Quick Links#
Posts and snippets on Python, machine learning, and neuroscience
Benchmark papers and academic publications
LLM evaluation, benchmarks, and open source libraries
Recent Posts#
16 December 2022 — TIL: Making a seaborn count plot with hue and labels
27 July 2022 — Deep auditory encoding model with self-attention to predict brain activity
13 November 2021 — Finding misspelled names with dirty_cat and unsupervised learning
31 July 2020 — An example workflow for voxel-wise encoding models using a BIDS app
23 January 2020 — Adding contours of a surface region to a statistical map in Nilearn