I'm a first-year MS Data Science student at UC San Diego and a Fulbright Scholar from Spain. I care about making AI systems more reliable and trustworthy, focusing on detecting AI-generated text, identifying hallucinations, and mitigating bias in language models. Before coming to UCSD, I spent two years as an AI Research Scientist at Accenture Labs in Dublin, where I filed patents and published papers on NLP and LLM evaluation. I've won several hackathons (HackUPC, Junction, AI Tinkerers Dublin) and love building things fast. When I'm not coding or writing papers, I'm probably thinking about transportation systems or giving talks about AI security.

AI security and robustness: watermarking, jailbreak detection, adversarial prompting. Causal inference for LLM evaluation. Multimodal models. Knowledge graphs + LLMs. Transportation AI and mobility optimization. Looking to collaborate on research projects, especially those bridging security, fairness, and practical deployment of LLMs.

Working on hallucination detection for SemEval 2025 (using perplexity-based approaches). Exploring Korean tonality as a potential attack vector for AI text detectors. Collaborating on gender bias mitigation through custom decoding strategies. Also building small tools to help with research workflows and always looking for new ideas at hackathons.