Mark Neumann
Principal Research Scientist at EvolutionaryScale
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolMark Neumann is a Principal Research Scientist with a decade of experience building ML systems across language, materials, and molecular science, currently leading protein and enzyme modeling at EvolutionaryScale. Previously a founding ML lead at Orbital Materials, he applied generative models to rational materials design, and before that helped develop foundational NLP tools at the Allen Institute (ELMo, AllenNLP, SciSpaCy). He combines research-grade representation learning with production-focused engineering, contributing model training, evaluation metrics, and robust pipelines for scientific domains. Based in Seattle, he pairs an MSc in Machine Learning from UCL and rigorous academic roots in mathematics/computer science with a knack for translating complex physical and biological problems into practical deep learning solutions. An under-the-radar strength is shipping reproducible tooling for domain-specialized NLP and NER—work that underpinned widely used projects like SciSpaCy.
10 years of coding experience
10 years of employment as a software developer
Natural Sciences Mathematics and Computer Science, Natural Sciences Mathematics and Computer Science at Durham UniversityUniversity College London
Github Skills (16)
bioinformatics10
pytorch10
machine-learning10
nlp10
spacy10
trainings10
python10
modeling10
data-science10
entities9
metric9
evaluation9
entity9
named-entity-recognition9
named-entity-extraction9
Programming languages (9)
Github contributions (5)
allenai/allennlp
Jun 2017 - Mar 2020
An open-source NLP research library, built on PyTorch.
Role in this project:
Data Scientist & Backend Developer
Data Scientist & Backend DeveloperContributions:2 releases, 309 commits, 785 PRs in 2 years 8 months
Contributions summary:Mark's commits primarily involve modifications to the AllenNLP library, focusing on the Semantic Role Labeling (SRL) model. They implemented and integrated a new dataset reader for the Winobias dataset, incorporated improvements like label smoothing and the handling of continued spans into the SRL model, and also added a new evaluation metric, EVALB. Their work indicates contributions to both the data preprocessing and the model implementation aspects of a machine learning project.
nlppytorchtransformerspythonbert
allenai/scispacy
Dec 2018 - Feb 2021
A full spaCy pipeline and models for scientific/biomedical documents.
Role in this project:Backend Developer & ML Engineer
Backend Developer & ML EngineerContributions:4 releases, 31 reviews, 106 commits in 2 years 2 months
Contributions summary:Mark primarily contributed to the training and evaluation of models within the SciSpaCy repository, focusing on the scientific and biomedical domain. Their work involved modifying and implementing scripts for retraining the parser and tagger, as well as developing code for training specialized NER models. Furthermore, the user made changes to evaluation metrics, including per-class scorers, indicating a focus on improving the model's performance and assessment. The user's efforts centered on enhancing the pipeline's capabilities for scientific document analysis.
nlppythonpipelinespacy-pipelinecustom-pipes
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.