William Roberts is a Berlin-based software developer with 15 years of experience building cloud-native backends, distributed systems, and production tooling. He has moved legacy, high-traffic e-commerce platforms to Kubernetes, designed microservice architectures and HA infrastructure (MySQL, Elasticsearch, Redis), and modernized frontend build pipelines using webpack and npm. His background blends academic rigor—a PhD in Computational Semantics and years teaching and research in NLP—with pragmatic engineering, evidenced by test automation contributions to widely used projects like NLTK. Comfortable across DevOps, full-stack and backend roles, he excels at turning research-grade language technology into reliable production services. Notably, he pairs deep language-modeling expertise with hands-on automation and deployment skills, making him adept at shipping scientifically informed, scalable systems.
15 years of coding experience
10 years of employment as a software developer
Doctor of Philosophy - PhD, Computational Semantics, Doctor of Philosophy - PhD, Computational Semantics at Humboldt University of Berlin
Bachelor of Science - BS, Cognitive Science, Bachelor of Science - BS, Cognitive Science at Queen's University
High School Diploma, High School Diploma at University of Toronto Schools
Master of Science - MS, Computational Linguistics, Master of Science - MS, Computational Linguistics at Universität des Saarlandes
Contributions:158 commits, 4 PRs, 11 comments in 2 years 2 months
Contributions summary:William's contributions focus on implementing and updating unit tests for the `nltk/nltk` repository, specifically targeting the `tgrep.py` module. The user wrote a series of unit tests to verify the tokenization process, various link types, and the functionality of tgrep searches. These tests cover diverse scenarios, including node matching, regular expression matching, dominance relations, and precedence relations, and also incorporated tests for handling of quoted node names and labeled nodes, ensuring the tgrep search feature's accuracy. Additionally, the user addressed encoding issues, verifying consistent behavior of tgrep search strings with both bytes and strings.
A flask template with Bootstrap, asset bundling+minification with webpack, starter templates, and registration/authentication. For use with cookiecutter.
Role in this project:
Full-stack Developer
Contributions:35 commits, 2 PRs, 6 comments in 7 months
Contributions summary:William primarily refactored the project to use modern frontend tooling. They replaced Bower and Flask-Assets with npm and webpack, integrating a build process for JavaScript and CSS. The user also updated the project's build process to include linting, asset management with file-loader, and environment variable configurations, modifying the testing and deployment pipelines. These changes demonstrate expertise in frontend build systems and deployment configurations.
pythonflaskasset-bundlingtemplateboilerplate
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.