David Speck
Senior AI Programmer at Guerrilla
Amsterdam, North Holland, Netherlands
Summary
👤
Senior🎓
Top SchoolDavid Speck is a Senior AI Programmer based in Amsterdam with 13 years of software engineering experience and a strong foundation in mathematics and formal AI training. He builds game-focused AI and production-ready ML systems, progressing from gameplay programming to senior AI roles at Guerrilla and engineering positions at Wooga and Aigent. His work bridges research and practice—evident in open-source contributions such as extending relational reasoning models in PyTorch to handle ternary relations and richer dataset generation. A former game-dev instructor, he combines clear pedagogy with hands-on delivery, making complex models practical for product teams. He often surfaces less obvious gains by improving tooling and observability (e.g., tensorboard integration) to speed iteration and debugging in ML workflows.
13 years of coding experience
13 years of employment as a software developer
Game Development, Game Development at S4G School for GamesMaster Artificial Intelligence, Master Artificial Intelligence at University of AmsterdamBachelor of Science (B.Sc.) Mathematik, Bachelor of Science (B.Sc.) Mathematik at Technische Universität BerlinBachelor of Science (B.Sc.) Mathematik, Bachelor of Science (B.Sc.) Mathematik at Durham University
German, English
Github Skills (11)
pytorch10
machine-learning10
deep-learning10
python10
tensorboard9
algorithms8
data-structures8
algorithm8
data-structure8
numpy7
computer-vision7
Programming languages (4)
Github contributions (5)
kimhc6028/relational-networks
Mar 2020 - Apr 2020
Pytorch implementation of "A simple neural network module for relational reasoning" (Relational Networks)
Role in this project:
ML Engineer
ML EngineerContributions:24 commits, 1 PR in 1 month
Contributions summary:David primarily contributed to the implementation and improvement of a relational reasoning model in PyTorch. Their work involved modifying the model architecture to handle ternary relations, enhancing the dataset generation process to include more complex relational questions (e.g., counting obtuse triangles), and integrating tensorboard logging for improved monitoring of training progress. They also added features for training on ternary relations, binary relations, and unary relations within the same training cycle.
pytorchrelationalsimple-neural-networkdeep-learningneural-network
Saduras/BattleSimGame
Aug 2020 - Sep 2022
Contributions:77 pushes, 3 branches in 2 years 1 month
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.