AI Safety Expert, English and Kannada

SaidGig; New York (NY); Teletrabalho; USD 22 000 - 30 000; Tempo integral

Vantagens oferecidas por esta oferta de emprego

Remote work

Resumo da oferta

SaidGig is seeking a text-based role to strengthen conversational AI systems by probing them with adversarial inputs, identifying weaknesses, and producing human-generated safety data.

You will red team models through jailbreaks, prompt injections, and misuse cases; review outputs on bias, misinformation, and harmful behavior; annotate failures and create reproducible reports and datasets for technical teams.

Qualificações

• Native fluency in English and Kannada.
• Clear written communication for technical and non-technical audiences.
• Ability to adapt across projects, task types, and customer needs.

Responsabilidades

• Red team conversational AI models and agents through jailbreaks, prompt injections, misuse cases, bias exploitation, and multi-turn manipulation.
• Review AI outputs involving sensitive subjects, including bias, misinformation, and harmful behaviors.
• Annotate failures, classify vulnerabilities, and flag systemic risks.
• Create reproducible reports, datasets, and attack cases that technical teams can use to address identified risks.

Conhecimentos

English

Kannada

Descrição da oferta de emprego

Role Overview

Help strengthen conversational AI systems by probing them with adversarial inputs, identifying weaknesses, and producing human-generated safety data. This text-based role focuses on testing AI behavior across challenging scenarios and turning findings into practical artifacts that improve system safety and reliability.

Key Responsibilities

• Red team conversational AI models and agents through jailbreaks, prompt injections, misuse cases, bias exploitation, and multi-turn manipulation.
• Review AI outputs involving sensitive subjects, including bias, misinformation, and harmful behaviors.
• Annotate failures, classify vulnerabilities, and flag systemic risks.
• Follow established taxonomies, benchmarks, playbooks, guidelines, and quality standards to conduct consistent testing.
• Create reproducible reports, datasets, and attack cases that technical teams can use to address identified risks.

Qualifications

• Native fluency in both English and Kannada is required.
• Strong judgment when assessing whether AI responses are accurate, complete, and appropriate, with the ability to explain your reasoning.
• Careful attention to subtle errors, inconsistencies, and gaps.
• Clear written communication for both technical and non-technical audiences.
• Ability to adapt across projects, task types, and customer needs.

Preferred Experience

• Adversarial machine learning, including jailbreak datasets, prompt injection, RLHF/DPO attacks, or model extraction.
• Cybersecurity experience, such as penetration testing, exploit development, or reverse engineering.
• Socio-technical risk work involving harassment or disinformation testing, abuse analysis or conversational AI testing.
• Creative adversarial thinking informed by psychology, acting, or writing.

Work Terms

• Remote, hourly engagement.
• All work is text-based.
• Participation in higher-sensitivity projects is optional. Topics are communicated before exposure, with clear guidelines and wellness resources available.

Compensation

16 to 22 per hour.