AI Studio Coach- LLM Red Teaming
Core
Coach undergraduate fellows in a 16-week program to ethically probe large language models for safety and security vulnerabilities using jailbreaking and prompt injection techniques.
Role type
Program Coach / Instructor (AI Safety & Red Teaming)
Builds
Knowledge, mindset, and skills for ethical LLM red teaming in a cohort of up to 500 fellows.
Domain
AI Safety, LLM Security, Red Teaming
Deliverable
production ML models
Required skills
LLM red teaming expertise, advanced prompt engineering, ethical red teaming practices, technical mentorship, event facilitation, constructive feedback, organizational skills
Preferred skills
Experience with Gray Swan Arena/Lakera Agent Breaker/HackAPrompt, background in AI security/trust and safety, teaching or mentoring experience
Technologies
LLMs, Canvas, Discord, Gray Swan Arena
Responsibilities
Facilitate workshops and maker days on red teaming concepts, provide technical mentorship on jailbreaking and prompt injection, grade assignments and provide feedback, monitor progress and intervene when needed, reinforce ethical practices
Seniority
Mid-level, hands-on IC