Overview
Freelance remote project to evaluate AI search responses for failures, working with Hindi-speaking experts on language evaluation.
What you'll do
- Identify realistic search-related failures where an AI model provides inadequate responses.
- Create search queries, test them, and document what went wrong.
- Write instructions and evaluation criteria for correct answers.
- Test multiple queries before finding failures.
What you'll need
- Research skills: ability to find and verify information from multiple sources, comfort navigating different websites, databases, and documentation.
- Critical thinking: can spot contradictions between sources, identify outdated information, and recognize when something doesn't apply to a specific locale or context.
- Attention to detail: can catch nuances in language, local variations (prices, regulations, requirements), and edge cases.
- Problem-solving mindset: comfortable iterating and testing when things don't work, and does not get frustrated by trial-and-error.
- Writing clarity: can articulate instructions and rubrics clearly so others understand the logic behind their reasoning.
Nice to have
- Domain knowledge: ideally a background in research, quality assurance, compliance, or similar work that requires precision and source verification.
- Locale knowledge: deep familiarity with their specific locale (local regulations, business practices, cultural nuances) is valuable for spotting region-specific failures.
Details
- Start date: ASAP.
- Duration: ongoing project.
- Commitment: 20 hours per week.
- Job type: freelance contract.
- Location: India, remote.
- Rate: $10 per hour.
Read the full description and apply on the company’s own careers page.