Summary
Pol Recasens is a PhD candidate in Computer Architecture based in Barcelona with eight years of experience bridging research and applied machine learning on high-performance computing platforms. He has held visiting researcher roles at IBM Research focused on LLM inference optimization and AI security—work that spans agentic red-teaming and production inference efficiency. His background includes hands-on contributions at Barcelona Supercomputing Center, CERN’s AtmoRep project, and DTU’s Cognitive Systems group, reflecting a mix of atmospheric deep learning, data science, and systems engineering. Pol moves comfortably between research and engineering, having shipped data-science solutions in industry and prototyped scalable ML models in research settings. He combines rigorous academic training from UPC, DTU and Politecnico di Milano with practical experience optimizing foundation models and inferencing pipelines. Colleagues describe him as driven by curiosity and collaboration, thriving on technically ambitious projects with passionate teams.
8 years of coding experience
2 years of employment as a software developer
Bachelor of Engineering - BE, Computer Science, Bachelor of Engineering - BE, Computer Science at Politecnico di Milano
UPC Universitat Politècnica de Catalunya
Technical University of Denmark
Catalan, English, Spanish