We are sharing a specialised remote opportunity for experienced AI red-teamers, cybersecurity specialists, or adversarial AI experts to support a leading AI lab in identifying vulnerabilities and improving the safety of advanced conversational AI systems.
This project focuses on probing AI models with adversarial prompts, identifying system weaknesses, and generating structured red-team datasets that strengthen the robustness and reliability of next-generation AI systems.
Key Responsibilities
Conduct adversarial testing of conversational AI systems
Identify vulnerabilities such as jailbreaks, prompt injections, misuse cases, and bias exploitation
Evaluate AI responses across multi-turn conversations and manipulation scenarios
Annotate model failures and classify vulnerability types
Generate structured reports, attack cases, and reproducible datasets
Follow testing frameworks, taxonomies, and benchmarks to ensure consistency
Produce clear documentation that helps teams strengthen AI safety systems
Ideal Profile
Strong candidates may have:
Native-level fluency in English and Hindi
Experience in AI red teaming, adversarial machine learning, or cybersecurity
Background in penetration testing, exploit development, or vulnerability analysis
Strong analytical thinking and structured problem-solving abilities
Ability to clearly explain risks to technical and non-technical audiences
Excellent written communication and attention to detail
Comfort working across multiple projects and testing scenarios
Nice to Have:
Experience with adversarial ML techniques (prompt injection, jailbreak testing, RLHF attacks)
Cybersecurity background in penetration testing or reverse engineering
Socio-technical risk analysis (bias, misinformation, abuse detection)
Creative adversarial thinking using psychology, narrative, or behavioral probing
Why This Opportunity
Contribute directly to improving the safety and robustness of advanced AI systems
Work on frontier AI red-teaming projects with leading AI research teams
Help identify vulnerabilities before AI systems reach production environments
Flexible remote project work with competitive compensation
Contract Details
Independent contractor role
Fully remote and flexible scheduling
Geographic eligibility restricted to India
Compensation: $13.87 per hour
Weekly payments via Stripe or Wise
Projects may extend or adjust depending on scope and performance
No access to confidential employer or institutional data required
About the Platform
This opportunity is available through a leading AI-driven work platform.
By submitting this application, you acknowledge that your information may be processed by 24-MAG LLC for recruitment and opportunity matching in accordance with our Privacy Policy:
Salary: $65,000 - 85,000 per year Requirements: ~ B.S. or B.A. in Computer Science or a related field preferred ~ Strong grasp of UI, full solution architecture, and .NET functions/standards ~ Familiarity with Oracle NetSuite and related integration solutions like...
...The Mary Sue: Assigning Editor The Mary Sue is looking for an experienced and highly motivated Assigning Editor to help develop and... ...opportunity to move into a full-time position after 36 months Fully remote Best suited to applicants based in the United States or the...
Position OverviewResponsible for providing service and technical support to our customers regarding technical issues or general inquires related to Sonosite ultrasound equipment.Company OverviewAt FUJIFILM Sonosite, we reinvent how healthcare is delivered with point-of-...
...Paralegal Admin Specialist Employment Type: Full-Time, Entry Level Department: Legal Support CGS is seeking a Paralegal to process electronic legal instruments, assist with trial prep and hearings, and provide additional legal support for a large federal agency...
...401(k) matching ~ Dental insurance ~ Health insurance ~ Vision insurance El Cajon Physical Therapy is looking for a Podiatrist to join our team! Our clinic is located in El Cajon, San Diego, close to Downtown San Diego. We are hiring a Podiatrist to diagnose...