Research Engineer, Value Persistence Through Reinforcement Learning
AI safety & policy
Full-time
AI safety & policy
Full-time
Salary
$200,000 to $300,000
Routes to impact
Skill-building & building career capital
Direct high impact on an important cause
Description
Run LLM training experiments on value persistence through synthetic midtraining, RL post-training, and evaluations with close collaboration from research scientists.
- Execute end-to-end training runs on open-weight models, mostly around the 12B scale
- Build and adapt RL pipelines, including RLVR with GRPO on verifiable tasks
- Create synthetic document corpora and matched control conditions for experiments
- Run evaluations on UK AISI's Inspect and interpretability methods under researcher guidance
- Debug training and infrastructure failures, and keep runs reproducible and well documented
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.



