Markus Kaiser
Research Manager at DeepL
Munich, Bavaria, Germany
Summary
👤
Senior🎓
Top SchoolMarkus Kaiser is a Research Manager and ML researcher with 13 years of experience bridging academia and industry, currently leading research at DeepL from Munich. He specializes in Physical AI—combining data generation, surrogate modelling, probabilistic methods (notably deep Gaussian processes and Bayesian neural nets), optimization, and scalable software to take projects from prototype to production. Over the last four years he has built and led multidisciplinary teams, defined ML and data strategy, and driven hiring and external collaborations for high-impact projects like rapid design search for stellarators and EV motor optimization. His hands-on open-source contributions include kernel work for GPflow and core parsing improvements in Holoviews, reflecting deep expertise in Gaussian processes and robust data visualization tooling. He has a PhD from TUM and a track record of publishing large physics-driven datasets (e.g., a 300k-design stellarator corpus) and securing collaborative research funding. Colleagues describe him as a practical probabilist who pairs rigorous Bayesian modelling with production-aware engineering.
13 years of coding experience
5 years of employment as a software developer
Master of Science - MS, Computer Science, Master of Science - MS, Computer Science at KTH Royal Institute of TechnologyDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Technical University of Munich
Github Skills (16)
gaussian-processes10
bayesian-statistics10
machine-learning10
bayesian10
holoviews10
parser10
tensorflow10
python10
plot10
gpflow10
kernel-methods10
testing10
pytest9
data-visualisation9
data-visualization9
Programming languages (9)
Github contributions (5)
GPflow/GPflow
Feb 2017 - May 2017
Gaussian processes in TensorFlow
Role in this project:
ML Engineer
ML EngineerContributions:18 commits, 10 PRs, 89 comments in 2 months
Contributions summary:Markus contributed to the implementation and testing of kernel functions within the GPflow library, a Gaussian process framework. Their work included the addition of a Multi-Layer Perceptron (MLP) kernel and the rework of the ArcCosine kernel, which involved defining functions and making the necessary changes to test files. They also optimized code, avoiding unnecessary data copies and ensuring correct parameter initialization.
information-theorygpflowdeep-learningmachine-learningmarkov-chain-monte-carlo
holoviz/holoviews
May 2016 - Dec 2016
With Holoviews, your data visualizes itself.
Role in this project:Back-end Developer
Back-end DeveloperContributions:9 commits, 3 PRs, 12 comments in 6 months
Contributions summary:Markus primarily focused on improving the Holoviews plotting library by modifying its internal parsing and option handling mechanisms. Their contributions include implementing and testing features for grouping paths without options, merging option dictionaries, and adding precedence rules for merging separate option definitions. The user also added tests to validate the correct behavior of combined options, highlighting a focus on ensuring the library's flexibility and reliability. The commits demonstrate expertise in modifying and extending core functionality related to parsing options.
visualizesholovizholoviewsplotting
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.