AI Quality Assurance Specialist — Remote Contract

Chagible; United States; Remote; USD 34,000 - 83,000; Part time

Job summary

Chagible is seeking a remote, independent contractor to rigorously test AI-powered applications. You will evaluate prompts, model responses, and multi-turn conversations, identifying inaccuracies, bias, and safety concerns.

Collaboration with engineers and prompt designers ensures fixes before and after release. You will document issues with reproducible steps, review training data quality, and support API integrations and performance monitoring.

Qualifications

• Hands-on experience testing generative AI, language models, or conversational apps.
• Familiarity with prompt and output evaluation, including multi-turn interactions.
• Ability to spot inaccurate, inconsistent, biased, or unsafe AI responses.
• Experience working with AI service APIs or model integrations and reviewing logs or performance indicators.
• Basic understanding of natural language processing and language model behavior.
• Careful, systematic approach to testing, edge cases, and issue documentation.

Responsibilities

• Run structured tests across prompts, response flows, and unusual or high-risk scenarios.
• Assess responses for accuracy, consistency, relevance, bias, hallucinations, and safety concerns.
• Check application behavior before and after deployment, including errors, logs, response stability, latency, and token use.
• Record findings with the relevant prompts, outputs, and clear steps for reproducing each issue.
• Review the quality and consistency of AI-generated training data and annotations.
• Collaborate with developers and prompt engineers to validate fixes and improvements.

Skills

AI testing

Prompt evaluation

Bias and safety spotting

APIs and logging review

NLP basics

Documentation

Job description

Chagible is seeking a remote, independent contractor to rigorously test AI-powered applications. You will evaluate prompts, model responses, and multi-turn conversations, identifying inaccuracies, bias, and safety concerns.

Collaboration with engineers and prompt designers ensures fixes before and after release. You will document issues with reproducible steps, review training data quality, and support API integrations and performance monitoring.