PhD Position in Mechanistic Interpretability: The University of Amsterdam is offering a PhD position focused on mechanistic interpretability of machine learning models. This research aims to enhance our understanding of how various deep learning architectures make predictions and contribute to making AI systems interpretable, reliable, and safe.
Designation:
PhD Candidate
Details:
| Field | Information |
|---|---|
| Research Area | Mechanistic interpretability in Machine Learning |
| Location | University of Amsterdam, Netherlands |
| Eligibility/Qualification | – Master’s degree in AI, Computer Science, Engineering, Mathematics, Physics, or a related discipline – Demonstrable background in Machine Learning – Excellent Python software engineering skills – Fluency in English (written and spoken) |
| Salary | € 2,872 to € 3,670 per month (scale P) |
| Contract Duration | 4 years (initially 18 months with extension) |
Description:
As a PhD candidate, you will engage in independent research on mechanistic interpretability. The candidate will focus on developing and evaluating post-hoc interpretability techniques, working with various types of AI models (e.g., transformers, GNNs) and various applications. Key responsibilities will include:
- Developing novel techniques to analyze information flow in deep neural networks.
- Creating evaluation frameworks for mechanistic interpretability techniques.
- Connecting empirical findings about model behavior with theoretical computation frameworks.
- Contributing to research publications for leading international AI conferences.
How to Apply:
Interested candidates should submit their applications online, including:
- A letter of motivation detailing research interests and reasons for applying (maximum 1 page)
- A list of Master-level modules taken with an official transcript
- A writing sample (e.g., Master’s thesis, term paper, or publication)
- A detailed CV with education and work experience
- A link to a GitHub repository, portfolio website, or relevant projects
Please compile the above documents into a single PDF for submission.
Last Date for Application:
Applications will be accepted until 31 December 2024.
For more details or inquiries, please contact Dr. Ana Lucic at a.lucic@uva.nl, quoting “PhD Position” in the email subject.
