Conduct uncertainty quantification research for large language models at the National Research Council in Ottawa. Apply deep learning and AI research skills.
This two-year Postdoctoral Fellow appointment focuses on calibration methods for LLM outputs and intermediate representations. You will investigate certainty metrics, interpretability, and trustworthy AI applications. The work supports research in Computational Linguistics and Machine Learning, with opportunities for publications and presentations.
Key Responsibilities:
- Develop uncertainty quantification methods for LLM generations
- Investigate uncertainty in intermediate model representations
- Assess signals for improved text-generation uncertainty metrics
- Apply methods to hallucination and deception detection
- Present research at competitive international venues
Requirements:
- PhD in Computer Science, Mathematics, Statistics, Physics, or related discipline
- Novel AI research project design and implementation experience
- Experience with deep learning tools and libraries
- Publication record showing experimental and scientific contributions
- PhD obtained within three years or completion expected within six months
Advance LLM uncertainty quantification through deep learning, calibration, interpretability, and rigorous computational research at NRC.