Aaron Klein
Group Lead - OpenEuroLLM at ELLIS Institute Tübingen
Freiburg im Breisgau, Baden-Württemberg, Germany
Summary
🤩
Rockstar🎓
Top SchoolAaron Klein is a machine learning leader with 12 years of experience who currently heads the OpenEuroLLM group at the ELLIS Institute, combining academic rigor from a PhD in machine learning with hands-on industry practice. He has led research teams and held scientist roles at AWS, contributed to prominent AutoML and Bayesian optimization open-source projects (auto-sklearn, SMAC3, RoBO, HpBandSter), and implemented practical HPO integrations such as LCNet wrappers and RandomForest surrogates. Comfortable bridging research and production, Aaron’s work spans Bayesian optimization, surrogate modeling, and automated model selection—areas he strengthened with unit-tested, production-aware code and runnable examples. He’s based in Freiburg, Germany, and known for advancing reproducible AutoML tooling while moving algorithms toward scalable, multithreaded deployments. An understated strength is his mix of leadership and deep implementation skill: he both structures optimization frameworks and writes the integration glue that makes them usable in practice.
12 years of coding experience
9 years of employment as a software developer
Dr. rer. nat. (Ph.D.), Computer Science, Machine Learning, Dr. rer. nat. (Ph.D.), Computer Science, Machine Learning at Albert-Ludwigs-Universität Freiburg im Breisgau
German, English, Spanish
Github Skills (31)
pytorch10
python10
experimental-design10
scikit10
machine-learning10
numpy10
hyperparameter-optimization10
bayesian10
automl10
optmization10
optimisation10
scikit-learn10
automated-machine-learning10
emulation10
random-forest10
Programming languages (5)
Github contributions (5)
automl/RoBO
Mar 2015 - Apr 2019
RoBO: a Robust Bayesian Optimization framework
Role in this project:
Backend Developer and Data Scientist
Backend Developer and Data ScientistContributions:786 commits, 70 PRs, 227 pushes in 4 years 2 months
Contributions summary:Aaron's commits focused on enhancing the RoBO framework, specifically related to surrogate model development. They implemented Lenet-on-MNIST surrogate tasks and incorporated random forest models for more efficient surrogate modeling. These changes involved modifications to the core Bayesian optimization solver, as well as adjustments to the acquisition function logic to support the use of different model types. The user also worked on providing unit tests.
bayesian
automl/HpBandSter
Dec 2017 - Mar 2019
a distributed Hyperband implementation on Steroids
Role in this project:ML Engineer
ML EngineerContributions:9 commits, 2 PRs, 5 pushes in 1 year 3 months
Contributions summary:Aaron contributed to the implementation and improvement of an LCNet wrapper within the hpbandster framework. Their work involved integrating LCNet for hyperparameter optimization, as evidenced by the addition of the `lcnet.py` file and modifications to the example usage. The user added threading capabilities to the LCNet wrapper. Furthermore, they integrated the `lcnet.py` into a runnable example for demonstrating how the LCNet could be utilized for the hyperparameter optimization process.
automated-machine-learningneural-architecture-searchhyperbandmultiobjective-optimizationhyperparameters
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.