Hi there! I'm Nawal Ould Amer. I work on making AI systems trustworthy. Currently a Research Scientist at Airbus, focused on uncertainty quantification and reliability for LLM-based systems.
I'm particularly interested in the gap between what looks good on a benchmark and what you can actually trust in a safety-critical environment. Lately I've been exploring model uncertainty from the inside, through mechanistic interpretability.
Interests
- Information retrieval and retrieval-augmented generation
- Uncertainty quantification and calibration for LLMs
- Reliability and evaluation of AI systems under real-world constraints
- Mechanistic interpretability — understanding model behavior from the inside
Outside of work
I’m building a personal multi-agent assistant, currently just for myself, running locally, that helps me track and connect things I’d otherwise forget (food, training, ideas, … etc). It’s also my playground for applying uncertainty quantification ideas to a real system: when should an agent ask for clarification vs just log its best guess? I’m planning to open it up for others to test once it’s further along.
Get in touch
The best way to reach me is by email: noa@nawalouldamer.com
The important things in life
| Paladin | Yuki |
|---|---|
 |
 |
Two cats. Very senior colleagues.