As a research engineer, you will create evals that test for misaligned model
behaviors. The goal will be to assess when models go off the rails, such as by
acting sycophantic, being dishonest, or taking unsafe actions.
We do not review applications submitted on Wellfound. Please apply here:
[https://mechanize.work/apply/research-engineer-
alignment/?utm_source=wellfound&utm_campaign=wellfound](https://mechanize.work/apply/research-
engineer-alignment/?utm_source=wellfound&utm_campaign=wellfound)
Applying takes less than two minutes.