HiringHigh payCreativeFast-growing
LLM Red Teamer (Safety Tester)
Aegis Research · Remote (Global) · Contract · Remote
$35 – $55 / hr
Why this role holds up next to AI
Finding the clever way a model can be tricked takes human creativity and mischief. That's not something AI is good at doing to itself.
What the work is
Try to break AI systems — get them to say something harmful or wrong — so they can be fixed before launch. Curiosity is the core skill.
Day to day, you'd
- Probe models for unsafe, biased, or false outputs
- Document repeatable ways to trigger failures
- Suggest guardrails the safety team can build
You're more ready than you think
You already do this if you're curious, skeptical, and like poking holes in things.
Apply for this role
Create your free account to apply — it takes under a minute.