Test conversational AI in English and Punjabi by uncovering jailbreaks, prompt injections, bias, and harmful behaviors. This part-time contract pays $16 to $22 per hour for 20 or more hours weekly.
About OpenTrain
OpenTrain AI is the hiring and contracting organization for this role. OpenTrain is the #1 platform for finding and building careers in AI training and data labeling, helping people discover projects, build a professional profile, and apply for AI work in minutes.
Creating an OpenTrain account is free, and your profile can help you present relevant experience as you grow in this rapidly expanding field.
About AI Safety Training
AI training is the human side of building artificial intelligence. Contributors evaluate model responses, identify risks, and prepare structured feedback that helps AI systems become more accurate, useful, and safe.
Red teaming is a specialized form of AI evaluation. It involves deliberately probing models with challenging or adversarial inputs so human reviewers can find weaknesses that automated testing may overlook.
- Remote work using text-based conversational AI evaluation
- Flexible, part-time contractor opportunity
- Direct contribution to the safety of modern AI systems
The Role
OpenTrain AI is recruiting a Punjabi Bilingual AI Safety Red Team Expert to test conversational AI models and agents with adversarial inputs. You will investigate jailbreaks, prompt injections, misuse cases, bias exploitation, and multi-turn manipulation.
The role includes reviewing outputs related to bias, misinformation, and harmful behaviors. Participation in higher-sensitivity projects is optional, with clear guidelines and wellness resources available.
- Pay: $16 to $22 USD per hour
- Time requirement: 20 or more hours per week
- Engagement: Part-time contractor
- Experience level: Entry level
- Languages: Native English and Punjabi
- Data type: Text
What You'll Do
You will turn adversarial testing findings into structured human data that supports model safety evaluation and actionable improvements. Consistent use of taxonomies, benchmarks, playbooks, and quality standards is central to the work.
- Test conversational AI models and agents for jailbreaks and prompt injections
- Probe misuse cases, bias exploitation, and multi-turn manipulation
- Review responses for accuracy, completeness, appropriateness, bias, and safety concerns
- Annotate model failures and classify vulnerabilities
- Identify systemic risks and weaknesses that automated testing may miss
- Document reproducible attack cases, reports, and datasets
- Apply evaluation guidelines, benchmarks, taxonomies, and quality standards consistently
Requirements
Native fluency in both English and Punjabi is required. You should have strong judgment when evaluating language, content, accuracy, completeness, appropriateness, bias, and harmful behavior.
Careful attention to subtle errors, inconsistencies, and gaps is important. You must be able to explain your reasoning clearly to both technical and non-technical audiences while following structured evaluation guidance.
- Native fluency in English and Punjabi
- Ability to identify jailbreaks, prompt injections, misuse cases, and multi-turn manipulation
- Ability to evaluate AI responses for accuracy, completeness, appropriateness, bias, and harmful behavior
- Ability to classify vulnerabilities and document reproducible attack cases
- Ability to follow taxonomies, benchmarks, playbooks, and quality standards
- Strong attention to detail and clear written reasoning
Helpful Background
Prior industry experience is not required for this entry-level opportunity, but relevant knowledge can help you approach adversarial testing from different perspectives. Experience in technical or behavioral analysis may be especially useful.
- Adversarial machine learning
- Cybersecurity or penetration testing
- Exploit development or reverse engineering
- Socio-technical risk or abuse analysis
- Conversational AI testing
- Psychology, acting, or writing
Location and Application
This opportunity is available to contractors located in Austria, Belgium, Bulgaria, Canada, Cyprus, Czechia, Germany, Denmark, Estonia, Spain, Finland, France, the United Kingdom, Greece, Croatia, Hungary, Ireland, Italy, Lithuania, Luxembourg, Latvia, Malta, the Netherlands, Poland, Portugal, Romania, Sweden, Slovenia, Slovakia, or the United States.
Apply through OpenTrain AI to be considered for this Punjabi bilingual red teaming opportunity. OpenTrain helps contributors find relevant AI training work, showcase credible experience, and build a lasting portfolio in the field.
- Work remotely with a computer and internet connection
- Choose a part-time schedule of 20 or more hours weekly
- Apply structured human judgment to cutting-edge AI safety evaluation