Marcel Kempers
Tech Lead - Senior Software Engineer at Starion
Amsterdam, North Holland, Netherlands
Summary
👤
Senior🎓
Top SchoolMarcel Kempers is a tech lead and senior software engineer based in Amsterdam with over 15 years of hands-on experience translating complex problems into production-ready solutions across finance, enterprise, and startups. He blends solution architecture with practical engineering, shipping cloud-native services in Azure (.NET Core, C#, Azure Functions, Azure SQL) while exploring applied ML and reinforcement learning with TensorFlow and contributions to the Tensorforce library. Marcel has led teams at banks and analytics firms, delivered innovative SharePoint/ERP integrations, and run a long-running freelance practice that sharpened his product-minded engineering instincts. He’s motivated by bringing good ideas to life and continuously evaluates emerging technologies for real-world impact. Beyond leadership, his open-source work improving preprocessors and RL examples shows a willingness to tackle niche technical challenges that make libraries more usable with image and grid-based state spaces. Colleagues describe him as a pragmatic architect who balances creativity, engineering rigor, and a clear focus on deployable outcomes.
9 years of coding experience
15 years of employment as a software developer
Master's degree Science and policy, Master's degree Science and policy at Utrecht University
English, Dutch
Github Skills (13)
deep-reinforcement-learning10
tensorflow10
python10
reinforcement-learning10
data-structure9
machine-learning9
algorithm9
data-structures9
algorithms9
numpy8
game-development7
mask-rcnn7
faster-rcnn7
Programming languages (3)
Github contributions (5)
tensorforce/tensorforce
Feb 2018 - Mar 2018
Tensorforce: a TensorFlow library for applied reinforcement learning
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 3 PRs, 3 comments in 1 month
Contributions summary:Marcel primarily contributed to the development of reinforcement learning models within the Tensorforce library. Their work included bug fixes in the maze explorer example and the addition of preprocessing layers, such as flatten, increase dimension, and expand dims, which enhance the ability of the library to be used with different state spaces, including image-based input and 2d matrices. The user also added the game 2048 and improved the behavior of preprocessors.
reinforcement-learningtensorflowdeep-reinforcement-learningtensorflow-librarytensorforce
mkempers/howto-sqlazure-pandas
May 2017 - Jan 2020
How to read and write to an Azure SQL database from a Pandas dataframe
Contributions:2 commits, 1 PR, 4 pushes in 2 years 9 months
azure-sql-databasepandas-dataframe
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.