Thomas Kober
Senior Applied Scientist at Microsoft
Berlin, Germany
Summary
🤩
Rockstar🎓
Top SchoolThomas Kober is a Senior Applied Scientist with 11 years of experience bridging NLP research and production-grade ML engineering, currently at Microsoft after leading applied-science efforts at Zalando. He holds a PhD in Computational Linguistics and has driven conversational AI work at Rasa—contributing to the popular open-source Rasa framework by improving its DIETClassifier, model checkpointing, and entity extraction. Equally at home in research and engineering, he progressed from postdoctoral work at Edinburgh to principal-level responsibilities, delivering robust NLP models and backend systems. Based in Berlin, he combines deep academic grounding with practical contributions to widely used open-source tooling, showcasing a knack for improving reliability and observability in ML pipelines.
11 years of coding experience
10 years of employment as a software developer
Doctor of Philosophy - PhD, Computational Linguistics, Doctor of Philosophy - PhD, Computational Linguistics at University of Sussex
English, Italian, French, German
Stackoverflow
5,900reputation
651kreached
62answers
6questions
neural-network
top-5%
hashset
top-5%
machine-learning
top-5%
nltk
top-1%
scikit-learn
top-5%
python
top-5%
Github Skills (26)
entity-extraction10
python10
nltk10
machine-learning10
tensorflow10
rasa10
natural-language-processing10
nlp10
model-checking9
checkpoint9
kmeans9
checkpointing9
scikit-learn9
neural-network9
hashset9
Programming languages (4)
Github contributions (5)
RasaHQ/rasa
Mar 2020 - Sep 2021
đź’¬ Open source machine learning framework to automate text- and voice-based conversations: NLU, dialogue management, connect to Slack, Facebook, and more - Create chatbots and voice assistants
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:123 reviews, 422 commits, 36 PRs in 1 year 6 months
Contributions summary:Thomas primarily focused on modifying and improving the DIETClassifier component within the Rasa framework. Their contributions involved adding class names to log messages for improved debugging, incorporating model checkpointing during training for better model persistence, and implementing configurable entity splitting by commas for improved entity extraction. The user also addressed issues in the logging of the component and updated the minimum compatible version of the Rasa framework to ensure that the latest Tensorboard logging feature works properly.
nlupythonbotspeech-recognitionbotkit
tttthomasssss/reversibledict
Apr 2017 - May 2023
Contributions:1 release, 9 pushes, 1 tag in 6 years 1 month
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.