Design and build secure, sandboxed evaluation environments for AI models to execute code and perform complex tasks. Develop backend infrastructure and agent scaffolding to support large-scale, repeatable AI evaluations and multi-step workflows.
The Cybersecurity Engineer will build secure infrastructure and automation tools to support adversarial testing and AI security evaluations. They will collaborate with red teamers to reproduce vulnerabilities and develop attack simulations to improve system resilience.
The Finance Manager will own and improve financial planning, accounting operations, and strategic finance capabilities. They will partner with leadership to provide actionable insights, manage budgets, and ensure accurate financial reporting.
Detect and investigate malicious cyber abuse, including ransomware and exploits, while conducting organizational-level threat actor analysis. The role involves querying internal data and OSINT to document findings and guide junior investigators on complex cases.
Review and label AI-generated exchanges to identify offensive cyber uplift and policy violations. Conduct investigations using SQL, Python, and OSINT to document findings and respond to security escalations.
Conduct adversarial testing and red teaming exercises to identify vulnerabilities and failure modes in advanced AI systems. Develop attack strategies, analyze model behavior, and document findings in actionable reports to improve AI safety.
Design, train, and deploy machine learning models across multimodal domains for safety, security, and intelligence applications. Build agentic systems and evaluation pipelines to measure model reliability and safety in production environments.
Design and execute red team assessments to evaluate the security and resilience of advanced AI systems and infrastructure. Identify vulnerabilities and failure modes to provide actionable recommendations for improving system safety.
Fellows will apply technical skills to high-impact research problems, contributing across the project lifecycle from data processing and visualization to model deployment and cloud infrastructure building. Responsibilities include collaborating on client-facing products, assisting with abuse detection automation research, and ideating new approaches in Trust & Safety and AI Security.
As a Threat Investigator, you will detect and respond to malicious use on the platform, investigate activities, draft reports, and make recommendations. You will also handle escalations and reactive leads that are not caught by existing safety systems.
As a Protection Scientist Engineer, you will design and build systems to proactively identify and enforce abuse on AI products. This includes monitoring new product launches and sustaining operations for existing products.
