Summary
Piotr Tarasiewicz is a Senior Software Engineer specializing in deep learning, probabilistic modelling, reinforcement learning and NLP, with a decade of experience building production-ready ML systems. Currently at NVIDIA, he develops reference DL solutions and optimizes GPU implementations, translating research methods into high-performance engineering. His background blends academic rigor (UCL ML MSc, work under Marc Deisenroth) with industry impact—shipping CTR/IPM models, probabilistic RTB predictors and RL agents for real-world ad and trading systems. Notable past projects include a 30% out-of-stock prediction model for a major e-commerce client and a prize-winning satellite image segmentation solution, highlighting his ability to move from novel algorithms to measurable business outcomes. Based in Warsaw, he pairs strong research instincts with hands-on engineering to deliver data-efficient, scalable ML products.
10 years of coding experience
3 years of employment as a software developer
Warsaw University of Technology
University College London
German, English, Polish