AI Safety & Reliability Analyst

ixolabs.ai; France; À distance; EUR 32 000 - 52 000; Temps partiel

Résumé du poste

ixolabs.ai is seeking an AI Safety Analyst to rigorously test and evaluate AI models, identify risks, and guide their development toward reliable and ethical outcomes. You will design comprehensive safety tests, craft adversarial prompts, and document failures such as hallucinations or biases.

This role emphasizes collaboration with red teamers and ethics reviewers, staying current with AI safety research, and providing actionable feedback to improve model robustness and fairness.

Qualifications

• Strong analytical and critical thinking skills.
• Experience with prompt engineering and interacting with large language models (LLMs) or generative AI systems.
• Familiarity with AI ethics, bias detection, and fairness metrics.
• Ability to think creatively and anticipate unexpected failure modes in complex systems.
• Excellent written communication skills for documenting findings and recommendations.

Responsabilités

• Design and execute comprehensive safety tests to identify harmful outputs, biases, and unintended behaviors in AI models.
• Develop adversarial prompts and scenarios to probe AI systems for vulnerabilities related to misinformation, privacy, and security.
• Categorize and document observed safety failures, including hallucination, toxicity, and discriminatory content generation.
• Provide detailed, actionable feedback to AI developers on improving model robustness, alignment, and ethical performance.
• Stay abreast of the latest research and methodologies in AI safety, interpretability, and responsible AI development.
• Collaborate with red teamers and ethics reviewers to create a holistic safety assessment framework.

Connaissances

Analytical thinking

Prompt engineering

LLM interaction

Ethics & bias awareness

Creative problem solving

Written communication

Formation

Background in CS or related field

Description du poste

ixolabs.ai is seeking an AI Safety Analyst to rigorously test and evaluate AI models, identify risks, and guide their development toward reliable and ethical outcomes. You will design comprehensive safety tests, craft adversarial prompts, and document failures such as hallucinations or biases.

This role emphasizes collaboration with red teamers and ethics reviewers, staying current with AI safety research, and providing actionable feedback to improve model robustness and fairness.