What the work actually is
- Model-output review: grade model answers on contracts, case analysis and memos against your own expertise.
- Task authoring: write problems a model can't fake — jurisdiction-specific traps, conflicting authority.
- Rubric design: define what correct legal reasoning looks like so a machine can be scored against it.
- Red-teaming: probe where legal models give dangerously wrong answers.
Who we look for
- JD or equivalent legal training and real practice experience.
- Bar admission helps but isn't required for every task type.
- Our network spans the US, East Africa, South Asia and more — US-only or EU-only project scoping is common, so say where you're licensed when you apply.
How it works
- Apply at experts.rustlabs.ai — it takes a few minutes.
- We review your background; when a legal project opens, we match people by field and results.
- Accepted work is paid at the project's rate. Identity verification and tax forms are completed before your first payout.