Frontier AI Safety Researcher: Detecting Capabilities

Independent; Greater London; Hybride; GBP 51 000 - 59 000; Plein temps

Avantages offerts par ce poste

43 days off per year

Generous pension schemes

World-leading institution

Résumé du poste

Imperial College London invites applications for a Postdoctoral Research Associate in the Safety & Alignment lab led by Dr. Oana-Maria Camburu. The funded role focuses on detecting unverbalized dangerous capabilities in frontier models and monitoring chain-of-thought reasoning.

You will design algorithms, conduct independent and collaborative research, mentor PhD/MSc students, and publish in NeurIPS, ICML, ICLR, ACL and related venues. The position starts ASAP for 12 months.

Qualifications

• PhD in CS/Math/Physics or related discipline or equivalent research/industrial experience.
• Publications in refereed journals/conferences (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP, TACL, AIJ).
• Experience with frontier models and deep learning libraries such as PyTorch.

Responsabilités

• Design and develop algorithms for detecting unverbalized dangerous capabilities in frontier models.
• Conduct independent and collaborative research within the group.
• Publish results in top-tier conferences and journals (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP, AIJ, TACL).
• Implement research ideas.
• Contribute to research projects involving frontier models.
• Participate in lab activities such weekly meetings and presentations.
• Mentor PhD and MSc students.

Connaissances

Research skills

Model checking

Monitoring

CoT reasoning

PyTorch

Formation

PhD in Computer Science, Mathematics, Physics or related

Outils

PyTorch

Description du poste

Imperial College London invites applications for a Postdoctoral Research Associate in the Safety & Alignment lab led by Dr. Oana-Maria Camburu. The funded role focuses on detecting unverbalized dangerous capabilities in frontier models and monitoring chain-of-thought reasoning.

You will design algorithms, conduct independent and collaborative research, mentor PhD/MSc students, and publish in NeurIPS, ICML, ICLR, ACL and related venues. The position starts ASAP for 12 months.