Summary
Karol Jezierski is a Principal AI Engineer based in Warsaw with nine years of hands-on experience building production-grade ML and agentic AI systems across healthcare, finance, and enterprise. He combines deep Python expertise with big-data engineering to design agent orchestration, tool-calling datasets, and large-scale pretraining pipelines, having led agent registries and multi-agent orchestration projects at Bayer and Roche. Karol has practical MLOps chops—containerized LLM fine-tuning, Kubernetes deployments, CI/CD on GitLab, and observability for LLMs—enabling teams to move research into reliable production. He has driven NLP and computer-vision solutions (OCR, hate-speech detection, YOLO-based annotation pipelines) and innovated synthetic-data and persona-generation techniques to improve SFT and DPO training. Notably, he reconstructed MoA ensembling and large-scale dialogue distillation workflows for Polish LLMs and implemented distributed deduplication pipelines with Ray and NVIDIA NeMo Curator. Pragmatic and stakeholder-focused, he pairs domain-driven product work with recruiting and technical strategy to embed AI responsibly at scale.
8 years of coding experience
7 years of employment as a software developer
Data Science with Python, Data Science with Python at Sages/Kodolamacz
Master Engineer’s Degree, Master Engineer’s Degree at Warsaw University of Technology
English