Summary
Carsten Draschner is an Applied Scientist with eight years of experience at the intersection of large-scale ML research and production-grade GenAI systems, currently advancing foundation-model lifecycles for Amazon’s Nova. He combines a Ph.D. in machine learning with hands-on expertise in multi-billion-token data pipelines, distributed training, SFT/RL pipelines and rigorous, non-contaminated evaluation frameworks to push coding capabilities and alignment. Previously he led applied LLM teams and built privacy-preserving GenAI stacks and autonomous evaluation systems for enterprise customers across regulated industries in Europe. His academic work on scalable, ethical ML for knowledge graphs informed open-source contributions to the SANSA Stack and EU H2020 projects, reflecting a sustained interest in sustainable and trustworthy AI. Based in Berlin, he pairs research-pedigree rigor with pragmatic engineering—optimizing data mixes and model-based filtering to squeeze token efficiency and real-world performance from large models.
8 years of coding experience
7 years of employment as a software developer
Doktor rer. nat.(Ph.D.) Informatik / Computer Science, Informatik, Doktor rer. nat.(Ph.D.) Informatik / Computer Science, Informatik at The University of Bonn
English, French, German