AI Safety Expert - Red Team

mercormercor·Remote(United States)
Other

WFA Digital Insight

In the rapidly evolving landscape of AI, safety experts are in high demand, with a 25% growth in job postings in the last year. As AI systems become more integral to our lives, ensuring their safety and security is paramount. Mercor, a leader in connecting top talent with AI research labs, is at the forefront of this effort. With the increase in AI adoption, companies are looking for experts who can red team conversational AI models and identify potential vulnerabilities. This role stands out due to its focus on AI safety and the opportunity to work with leading AI research labs. Before applying, candidates should be aware of the importance of fluency in English and Malayalam, as well as prior experience in AI adversarial work or cybersecurity.

Job Description

About the Role

The AI Safety Expert - Red Team role at mercor is a unique opportunity to work at the intersection of AI, cybersecurity, and socio-technical probing. As an expert in this field, you will be responsible for testing the limits of conversational AI models and identifying potential vulnerabilities. Your work will contribute directly to the development of safer and more secure AI systems. The role is remote, offering the flexibility to work from anywhere, and is contract-based, providing the opportunity to work on a project basis.

The day-to-day responsibilities of this role will involve red teaming conversational AI models, performing jailbreaks, prompt injections, and misuse cases to identify vulnerabilities. You will also be responsible for generating high-quality human data, annotating failures, and classifying vulnerabilities. Your work will require strong communication skills, as you will need to explain complex risks to both technical and non-technical stakeholders.

The team you will be working with is comprised of experts in AI, cybersecurity, and socio-technical probing. You will have the opportunity to collaborate with leading AI research labs and contribute to the development of cutting-edge AI safety protocols. The reporting structure for this role is not specified, but you can expect to work closely with the mercor team and other stakeholders.

What You Will Do

  • Red team conversational AI models to identify vulnerabilities and weaknesses
  • Perform jailbreaks, prompt injections, and misuse cases to test AI model limits
  • Generate high-quality human data to inform AI model development
  • Annotate failures and classify vulnerabilities to improve AI model safety
  • Explain complex risks to technical and non-technical stakeholders
  • Use taxonomies, benchmarks, and playbooks to maintain consistent testing
  • Document findings and produce reports, datasets, and attack cases
  • Collaborate with leading AI research labs to develop cutting-edge AI safety protocols
  • Apply structure to testing using established methodologies

What We Are Looking For

  • Fluency in English and Malayalam
  • Prior experience in AI adversarial work, cybersecurity, or socio-technical probing
  • Strong communication skills for explaining risks to technical and non-technical stakeholders
  • Experience with Adversarial ML, Cybersecurity, and socio-technical risk analysis
  • Skills in creative probing, such as psychology, acting, or writing
  • Ability to work independently and collaboratively as part of a remote team
  • Strong problem-solving skills and attention to detail

Nice to Have

  • Experience with conversational AI models and their applications
  • Knowledge of AI ethics and safety protocols
  • Familiarity with red teaming methodologies and tools
  • Certification in AI safety or a related field

Benefits and Perks

  • Competitive hourly compensation
  • Opportunity to work with leading AI research labs
  • Collaborative and dynamic remote work environment
  • Flexible contract-based work arrangement
  • Professional development opportunities in AI safety and related fields
  • Access to cutting-edge AI technologies and tools
  • Recognition for contributions to AI safety and security

How to Stand Out

  • Tip: Make sure to highlight your experience with AI adversarial work, cybersecurity, or socio-technical probing in your application.
  • Tip: Practice explaining complex technical concepts in simple terms to demonstrate your communication skills.
  • Tip: Review the mercor website and understand their approach to AI safety and research labs before applying.
  • Tip: Be prepared to provide examples of your creative probing skills, such as writing or acting experience.
  • Tip: Research the current state of AI safety and security to demonstrate your knowledge and interest in the field.
  • Tip: Prepare to discuss your experience with taxonomies, benchmarks, and playbooks in the context of AI safety testing.

This is a remote position listed on WFA Digital, the platform for professionals who work from anywhere. Browse more remote jobs across all categories.