Paul Dubs
Chief Technical Officer at Paul Dubs IT Consulting
Mainz, Rhineland-Palatinate, Germany
Summary
🤩
Rockstar🎓
Top SchoolPaul Dubs is a seasoned technology leader and CTO with over a decade focused on machine learning and 13+ years of hands-on software experience, now co-founding and steering Xpress AI from Mainz. He combines product and engineering leadership—from deep learning work on JVM-centric projects like deeplearning4j to building a graphical platform that lets customers design and customize AI agents for cloud or on-prem deployments. Paul’s background spans founding consultancies, senior R&D and developer support roles, and product leadership at Skymind, giving him a rare blend of low-level performance optimization and high-level product vision. He’s an active open-source contributor who has improved model import, data handling and performance in widely used ML tooling, signaling a pragmatic focus on production-ready AI. Notably, his platform emphasizes extensibility and user-driven customization, putting advanced AI capabilities directly in clients’ hands.
13 years of coding experience
7 years of employment as a software developer
Master of Science (MSc), Informatik, Master of Science (MSc), Informatik at Technische Universität Darmstadt / TU Darmstadt
English, French, Russian, German
Stackoverflow
807reputation
28kreached
35answers
0questions
Github Skills (16)
neural-network10
javas10
keras10
machine-learning10
deep-learning10
neural10
java10
optimization9
linear-algebra9
artificial-intelligence9
deeplearning4j6
nd4j6
dl4j6
python6
war6
Programming languages (12)
Github contributions (5)
deeplearning4j/deeplearning4j
Mar 2016 - Oct 2021
Suite of tools for deploying and training deep learning models using the JVM. Highlights include model import for keras, tensorflow, and onnx/pytorch, a modular and tiny c++ library for running math code and a java based math library on top of the core c++ library. Also includes samediff: a pytorch/tensorflow like library for running deep learn...
Role in this project:
Backend Developer
Backend DeveloperContributions:773 reviews, 194 commits, 59 PRs in 5 years 8 months
Contributions summary:Paul made several contributions focused on improving the core functionality of the deeplearning4j library. They optimized data handling within the LineRecordReader, added support for custom loss functions for Keras models, and refactored and enhanced UI elements. Furthermore, they removed the GloVe embedding, which demonstrates active maintenance and a focus on better-performing alternatives. These commits suggest a focus on performance, model import, and general library improvements.
deeplearning4jpythonautomatic-differentiationneural-netskeras-tensorflow
deeplearning4j/deeplearning4j-docs
Mar 2018 - Nov 2020
Documentation for Deeplearning4j - Deep Learning for the JVM, Java & Scala
Contributions:21 commits, 10 PRs, 16 pushes in 2 years 8 months
deeplearning4jdeep-learningmachine-learning-algorithmsmachinelearningmachine-learning
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.