Caner Türkmen
Senior Applied Scientist at Amazon Web Services (AWS)
Berlin, Germany
Summary
🤩
Rockstar🎓
Top SchoolCaner Türkmen is an applied scientist and former management consultant with 12 years of experience building and deploying advanced analytics and probabilistic ML solutions across seven markets on four continents. Based in Berlin and currently at AWS, he combines hands-on engineering—contributing to prominent open-source projects like GluonTS and AutoGluon—with a research focus on Bayesian inference and temporal point processes. He has a strong academic foundation (MSc and PhD work in engineering/computer science) and practical experience enabling production-ready forecasting models and hyperparameter-optimizable GluonTS integrations. An instructor as well as practitioner, he teaches analytics fundamentals and tools, translating complex probabilistic ideas into usable code and clear learning paths. Unusually for someone with consultancy roots, he is deeply technical in time-series and TPP modeling, having added core transforms and variable-length data support to widely used forecasting libraries.
12 years of coding experience
PhD, Computer Engineering, PhD, Computer Engineering at Boğaziçi UniversityBS, Management Engineering, BS, Management Engineering at Università degli Studi di Roma 'La Sapienza'BSc, Industrial Engineering, BSc, Industrial Engineering at Boğaziçi Üniversitesi
Stackoverflow
61reputation
63kreached
2answers
0questions
Github Skills (20)
pytorch10
python10
data-science10
machine-learning10
time-series10
forecasting10
hyperparameter-optimization10
deeplearning-ai10
time-series-forecasting10
deep-learning10
forecast10
automated-machine-learning10
artificial-neural-networks9
ensemble-learning9
mxnet9
Programming languages (4)
Github contributions (5)
autogluon/autogluon
Mar 2022 - Nov 2022
Fast and Accurate ML in 3 Lines of Code
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:583 reviews, 45 commits, 180 PRs in 8 months
Contributions summary:Caner made significant contributions to the forecasting models within the AutoGluon repository. Their work involved adding and refactoring model objects for forecasting, including the addition of new models like Prophet and a generic GluonTS model. The user also implemented several test cases for these models, contributing to the improvement of model performance and robustness. Furthermore, the user focused on enhancing the handling of constructor arguments for the GluonTS models, enabling hyperparameter optimization capabilities.
forecastingimage-textmlppythonmeta-learning
awslabs/gluonts
Dec 2019 - Mar 2022
Probabilistic time series modeling in Python
Role in this project:ML Engineer / Data Scientist
ML Engineer / Data ScientistContributions:24 reviews, 12 commits, 26 PRs in 2 years 4 months
Contributions summary:Caner's contributions primarily involve adding new transform objects related to temporal point processes and implementing label smoothing for the binned distribution. They modified existing code within the `gluonts` library, focused on time series modeling, indicating a focus on extending or enhancing the core functionality. The changes included code to support variable length data loading in preparation for the deep renewal processes. Furthermore, the user added SampleForecast and Predictor objects for TPPs along with adjustments to data loading.
forecastingpythontime-series-analysistimeseries-forecastingaws
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.