About T3

T3 is a specialist AI assurance and governance consultancy helping organisations deploy AI safely, responsibly and at scale. We work with leading technology companies, financial institutions and public sector organisations to design, evaluate and assure AI systems, combining deep technical expertise with practical implementation.

About the Role

We are seeking a Senior AI Safety Red Teamer with extensive experience working on frontier AI systems. This is an opportunity to apply your expertise across a diverse range of clients, helping organisations understand, evaluate and mitigate the risks associated with advanced AI models and AI-enabled products.

Key Responsibilities

- Lead AI safety red teaming engagements for frontier foundation models, LLMs and AI agents.
- Design and execute adversarial evaluations covering jailbreaks, prompt injection, data poisoning, harmful content, misuse, autonomous capabilities and emerging model behaviours.
- Develop novel attack methodologies to identify previously unknown failure modes.
- Build and enhance automated evaluation pipelines and safety benchmarks.
- Produce high-quality technical reports with clear, evidence-based findings and recommendations.
- Advise clients on AI safety testing strategies, evaluation methodologies and model assurance practices.
- Contribute to the development of T3’s AI safety and frontier AI assurance methodologies.
- Stay at the forefront of AI safety research, emerging threats and evaluation techniques.
- Support business development through thought leadership, client workshops and technical discussions.

Essential Requirements

- Extensive experience conducting AI safety red teaming for frontier AI systems.
- Previous experience working within,



or in close collaboration with, a leading AI laboratory (e.g. OpenAI, Anthropic, Google DeepMind, Meta, xAI, Microsoft AI, Amazon AGI or equivalent).
- Deep understanding of large language models, multimodal models, reasoning models and AI agents.
- Demonstrated expertise in adversarial testing, jailbreak development, prompt injection, model evaluation and misuse testing.
- Strong Python skills and experience developing automated AI evaluation tooling.
- Experience using AI evaluation frameworks such as Inspect AI, Garak, PyRIT, DeepEval or equivalent.
- Strong understanding of AI safety, model alignment, AI security and frontier AI risk research.
- Excellent communication skills, with the ability to explain complex technical concepts to both technical and executive audiences.

Desirable Experience

- Published AI safety research or recognised contributions to the AI safety community.
- Experience evaluating multimodal and agentic AI systems.
- Experience designing scalable evaluation frameworks and automated testing pipelines.
- Familiarity with emerging AI assurance standards and governance frameworks.
- Experience presenting findings to senior technical leadership and executive stakeholders.

Preferred Expertise Candidates should demonstrate expertise across several of the following areas:

- Frontier AI evaluation
- AI safety research
- LLM and agent red teaming
- Prompt injection and jailbreak testing
- Autonomous capability assessments
- Harmful content and misuse evaluations
- AI security testing
- Model alignment evaluation
- AI assurance
- Adversarial machine learning
- Safety benchmarking
- Continuous model monitoring
- AI risk assessment
- Evaluation methodology design

📌 Senior AI Safety Red Teamer (Argentina)
🏢 T3
📍 Argentina