FAR.AI
FAR.AI
FAR.AI is a research and education nonprofit working to make advanced artificial intelligence safe and beneficial for everyone. Its team studies key challenges in trustworthy AI, including robustness, deception, interpretability, and the evaluation of frontier systems. Through technical research, training, events, workshops, grantmaking, and programs such as FAR.Labs, FAR.AI supports researchers and strengthens the field’s capacity to develop secure and aligned AI. The organization also shares research findings, convenes academic and industry leaders, and collaborates with universities and public agencies on AI safety and policy.

Head of Engineering, AI Red Team

Lead FAR.AI’s engineering function for a global frontier-AI red-teaming program. Build teams, agentic tooling, and evaluation systems that advance AI safety and security.

Description

  • Scale FAR.AI’s Red Team engineering organization and its systems, products, services, evaluations, and agents
  • Create agentic red-teaming systems that systematically investigate attack surfaces
  • Discover newly disclosed public jailbreaks and incorporate them automatically into red-teaming platforms
  • Improve attacker simulation and harmful-output evaluation methods
  • Develop vulnerability databases, statistical analysis capabilities, and reporting infrastructure
  • Deliver systems for high-stakes red-teaming engagements with frontier AI companies and governments
  • Help produce public reports, benchmarks, and leaderboards
  • Participate in red-teaming both closed- and open-weight frontier models
  • Develop agentic workflows that increase red-team efficiency
  • Lead, mentor, and support engineering and red-teaming contributors
  • Plan sprint work and direct engineering delivery
  • Establish scalable processes for fast-moving engagements
  • Design hiring pipelines and recruit engineering managers and other team members
  • Find engineering constraints and put effective remedies in place
  • Work with technical leadership on engineering systems, methods, evaluations, and research
  • Help shape red-team technical and product strategy
  • Turn technical concepts and findings into practical real-world impact

Requirements

  • Substantial software engineering, AI, or computer-engineering experience, including areas such as cybersecurity or MLOps
  • A strong record of building, developing, and leading technical teams, including startups from inception and teams of managers
  • Hands-on experience with coding agents, Python, LLM APIs, local LLM frameworks, evaluation frameworks, LLM agents, cloud infrastructure, and compute clusters
  • Ability to perform effectively in rapidly changing environments
  • Clear commitment to advancing the mission and impact of frontier AI systems
  • Strong communication skills for explaining technical solutions to both technical and non-technical audiences
  • Persistent execution toward ambitious objectives
  • Experience building or red-teaming frontier LLMs or agentic systems is advantageous
  • Experience creating technical teams and products in entrepreneurial settings is advantageous
  • Ability to uncover subtle, high-severity vulnerabilities in complex systems is advantageous
  • Background in adversarial machine learning or security is advantageous
  • Previous work with AI labs, security teams, or government safety institutes is advantageous
  • Published research or other work in AI safety, security, or robustness is advantageous
  • Willingness to work with some timezone flexibility
  • Availability for full-time work of 40 hours per week

Benefits

  • Work-related travel costs are covered
  • Equipment costs are covered
  • Catered lunch and dinner are provided at the Berkeley office
  • Visa sponsorship is available for the US or Singapore
  • Paid work trial

Related Jobs