Summary
Joel Niklaus is a Data Lead and machine learning engineer at Hugging Face with 11 years of experience building and curating pretraining datasets and large language model systems, particularly for legal applications. He holds a PhD in NLP and has trained multi-billion-parameter models on hundreds of TPUs at X, achieving state-of-the-art results on LegalBench and helping establish Swiss legal NLP datasets used by courts and researchers. Joel’s work spans industry and academia—from Research Scientist at Harvey and Thomson Reuters Labs to visiting research at Stanford—and has earned an Outstanding Paper Award at ACL and media coverage from Anthropic and Swiss National Radio. He advises and angel-invests in AI startups, contributes to open-source tools like lighteval and Marin, and teaches NLP to students and professionals. Less obvious: he combines deep technical pretraining expertise with practical dataset engineering across diverse compute environments, making him equally fluent in research benchmarks and production-ready data pipelines.
11 years of coding experience
12 years of employment as a software developer
Bachelor of Science - BS Mathematics and Computer Science, Bachelor of Science - BS Mathematics and Computer Science at University of Exeter
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of Bern
Matura Classical language profile, Matura Classical language profile at Kantonsschule Rychenberg Winterthur
Master of Science - MS Computer Science, Master of Science - MS Computer Science at Nanyang Technological University Singapore
Visiting Student Researcher Legal Natural Language Processing, Visiting Student Researcher Legal Natural Language Processing at Stanford University
English, French, griechisch, alt (bis 1453), Latin, German, Spanish, Italian, Chinese