rlhf services talent and countries

What does this nearshore talent hub cover?

Use this collection to choose a role and build a country shortlist for your RLHF team. Start with the tasks the hire will own: ranking responses, checking code, designing evaluations or investigating safety failures.

Write down who will resolve disputed ratings. Ask candidates to work through an example where reasonable reviewers might disagree. Their explanation gives you something concrete to discuss in the interview.

The role guides help you frame that assessment. The country guides help you check working hours and prepare questions about the employment quote. Treat country selection as a scheduling and hiring decision. Test each person's skills directly.

Before you compare candidates, describe what a finished task looks like. For response ranking, include the prompt, candidate responses and scoring rules. For code review, include a small repository and the tests a correct solution must pass. Use sample data that you have permission to share.

Decide how much guidance the hire will receive. Someone applying an established rubric has a different assignment from someone writing that rubric. Put that distinction in the role description, along with the name of the manager who can answer questions during onboarding.

This site is an Opus-owned reference; recommendations are editorial.

Talk to Opus about a role.

Which entries should you read first?

Choose the guide closest to the work waiting in your queue:

• RLHF annotator and AI trainer: Start here when you need people to compare responses and apply a rating rubric. Prepare examples of acceptable ratings and cases that need escalation.
• Coding agent trainer: Use this guide for code review and programming tasks. Specify the languages, test tools and repository access the work requires.
• Model evaluation specialist: Read this when the hire will help define evaluations or analyze failures. Bring an example of the decision an evaluation must support.
• AI safety researcher: Start here for safety-focused work. Define the risk area and who approves testing before interviewing.

Then compare locations against your team's actual calendar:

These city offsets follow the IANA time zone records for North America and South America. Confirm the candidate's city and agreed hours, including what happens when US clocks change.

Write the required overlap into the role description. Include live review meetings and the hours when a manager can unblock work. Ask candidates to confirm that schedule before the final interview. For tasks that can run without live supervision, define the handoff and when feedback is due.

Before requesting a shortlist, complete a role scorecard and plan the interview loop. Name the person who will review the hire's work. If you need a supplier to own batch delivery, use the hiring models guide to prepare questions about supervision and acceptance criteria.

Send the same task description to each provider you assess. Request a written quote tied to the role, location and working schedule so you can compare the same scope. Ask which costs recur, what can change the rate and how leave affects coverage. Keep those answers beside the candidate's work sample when you make the hiring decision.

Talk to Opus about a role.

<!-- Not saved to disk: the workspace is read-only. Live links were not checked. -->