Applied Transformer NLP for Cybersecurity
A hands-on two-hour workshop covering applied transformer-based NLP for cybersecurity using the HuggingFace Transformers library. Participants build a complete threat intelligence pipeline from scratch — entity extraction, classification, summarization, and semantic ATT&CK mapping — using only open-source models running locally.
This workshop teaches cybersecurity professionals how to apply modern transformer models to real security problems using the HuggingFace Transformers library. The session covers the full spectrum of NLP tasks relevant to security operations — from understanding unstructured threat reports to automated technique mapping against the MITRE ATT&CK framework.
Participants gain practical experience choosing the right model architecture for the right problem, a critical skill as transformer-based tools become standard in threat intelligence, SOC automation, and security data analysis. The session closes with a live demonstration showing how individual components combine into a fully automated analysis pipeline.
Pauline Bourmeau is an independent security researcher specializing in the intersection of artificial intelligence, cognitive psychology, and threat intelligence. She has consulted on multilingual natural language processing, led deep learning and NLP workshops, and created training materials blending STEM with human factors. As founder of DEFCON Paris and contributor to the MISP project, she actively advances collaborative cybersecurity practices.
Previously, Pauline worked as a Threat Intelligence Analyst conducting OSINT, HUMINT, and SOCINT analysis to profile threats and investigate APTs. She holds a Master’s in Criminology with a thesis on cybersecurity intelligence sharing, and a background in sociolinguistics and computer science from Sorbonne and School 42.