Scholar
Francisco Eiras
Google Scholar ID: O_iJTgYAAAAJ
University of Oxford
Citations & Impact
Citations
479
H-index
10
i10-index
11
Publications
20
Co-authors
9
list available
Publications
Resume (English only)
Academic Achievements
- Know Thy Judge: On the Robustness Meta-Evaluation of LLM Safety Judges (ICLR Workshop Proceedings, 2025)
- Do as I do (Safely): Mitigating Task-Specific Fine-tuning Risks in Large Language Models (ICLR, 2025)
- Risks and Opportunities of Open-Source Generative AI (arXiv, 2024)
- Near to Mid-term Risks and Opportunities of Open-Source Generative AI (ICML Position Paper, 2024)
- Efficient Error Certification of Physics-Informed Neural Networks (ICML, 2024)
- Segment, Select, Correct: A Framework for Weakly-Supervised Referring Segmentation (ECCV Workshop Proceedings, 2024)
- Provably Correct Physics-Informed Neural Networks (ICML Workshop, 2023, Outstanding Paper Award)
Research Experience
- Dynamo AI, Nov. 2024 - Present: ML Research Scientist, leading the research team developing agentic red-team evaluations for LLMs, targeting custom safety specifications in conversational and tool-using agents, designing automated pipelines to increase adversarial coverage and behavioral diversity across evaluation scenarios, contributing to core research efforts, including a first-author paper on the robustness of LLM safety judges.
- Five, Dec. 2022 - Jun. 2023: Research Scientist Intern, worked on efficient methods to perform zero-shot and weakly-supervised referring image segmentation, achieving new state-of-the-art performance in the field.
- Five, Jun. 2021 - Sep. 2021: Research Scientist Intern, extended the certified robustness technique of randomized smoothing from isotropic ℓp balls to anisotropic certificates through a simplified Lipschitz analysis-based framework.
- Five, Sep. 2018 - Sep. 2020: Research Engineer, led the development of safe and scalable optimization-based motion planning algorithms, published and presented research work developed at top-tier conferences and journals within the robotics community, wrote and reviewed research and development code, ensuring CI with other tools within the company.
- Institute for Systems and Robotics, Lisbon, Apr. 2017 - Sep. 2017: Graduate Research Assistant, developed new methods to perform pose estimation using vanishing points in general (central and non-central) omnidirectional cameras, leading to a CVPR 2018 paper.
Education
- Ph.D. from the University of Oxford under the supervision of Prof. Philip Torr, Dr. Adel Bibi, and Dr. M. Pawan Kumar (Google DeepMind), working on trustworthy machine learning, with a focus on the safety and security of large language models.
Background
- ML Research Scientist focusing on compliance and safety evaluations of generative AI systems, with an emphasis on large language models.
Miscellany
- Personal interests and contact information not detailed.
