Job Summary
Johns Hopkins University’s Center for Language and Speech Processing (CLSP) is recruiting a Postdoctoral Fellow to advance the understanding of foundation models. The role focuses on mechanistic interpretability, multimodal AI, and developing methods for trustworthy and controllable artificial intelligence.
Job Details
| Designation | Postdoctoral Fellow |
| Research Area | Mechanistic Interpretability, LLMs, Multimodal AI |
| Location | Baltimore, MD |
| Salary | $80,000 (Budgeted) |
| Open Date | October 30, 2026 |
| Last Date for Apply | Not Specified (Open until filled) |
Research Area
The research focuses on the following domains:
- Mechanistic interpretability of large language models.
- Multimodal AI and machine learning.
- Representation analysis and causal interventions.
- Model steering and controllable AI.
- Speech, audio, or language modeling.
Eligibility/Qualification
- Candidates must hold a doctoral degree (Ph.D.) in Electrical Engineering, Computer Science, or a closely related field.
- Strong expertise in machine learning and mechanistic interpretability is essential.
Job Description
The successful candidate will be advised by Berrak Sisman and Philipp Koehn, collaborating with an international team at the Whiting School of Engineering. The fellow will work on high-impact research projects aimed at making foundation models more interpretable and trustworthy. Key responsibilities include conducting original research, performing representation analysis, and developing methods for causal interventions in AI models.
How to Apply
Interested candidates should submit the following documents via Interfolio:
- A current CV with a full list of publications.
- A brief description of research experience and interests regarding mechanistic interpretability.
- Contact information for two professional references.
For inquiries, please contact Berrak Sisman (sisman@jhu.edu) and Philipp Koehn (phi@jhu.edu).
Apply Link
You can submit your application through the official portal here: Apply Now






