AIML - Machine Learning Manager, Red Teaming - Responsible AI and Safety

Apples Responsible AI and Safety team focuses on innovative technologies, methodologies, and research to enable fantastic user experiences and to push the frontier of machine learning. Our team is looking to hire a leader with a strong track record in AI red-teaming and adversarial security research, who is passionate about proactively finding real vulnerabilities in generative AI systems before they reach users.In this role, you will lead a dedicated red-teaming team, building and running an independent program that actively probes Apple Intelligence features for safety and security risks, while upholding Apples values, privacy, and quality standards. This role focuses on leading a team that sets its own red-teaming agenda, independent from standard evaluation, and is responsible for actively probing Apple Intelligence features to discover real vulnerabilities rather than measuring known ones. This team is also responsible for building the operating model to support that work: hiring and developing red-teaming engineers with sustained feature ownership, working with a skill crowd workforce paired directly with those engineers, and standing up automated red-teaming capabilities.Leading this function requires clear division of responsibility with adjacent security and evaluation teams, sound judgment on which findings warrant escalation, and the ability to define success criteria for a program that doesnt yet have established metrics. This role also involves managing external red-teaming partnerships to extend coverage at scale, and communicating findings and program impact to leadership on a recurring basis.This role heavily draws on technical leadership in AI safety or security, program-building in ambiguous or newly-scoped territory, cross-functional negotiation, and metrics definition and reporting.Build and lead a team of red-teaming engineers, each with sustained ownership of a feature clusterSet an independent red-teaming agenda, distinct from standard evaluation, focused on active vulnerability discoveryOwn coverage strategy across pre-ship qualification, post-ship monitoring, and foundation model red-teaming, in partnership with relevant model teamsDefine division of responsibility with adjacent security teams on cross-cutting risks (e.g., prompt injection) and coordinate escalation pathsOwn external red-teaming partnerships, including scoping value, cost, and pipeline integrationDefine, instrument, and report on red-teaming success criteria to leadership on a recurring basisProvide technical direction on automated red-teaming capabilitiesWork with highly sensitive content, including exposure to offensive and controversial materialBuild and report high visibility and compliant reporting capabilitiesMS, or PhD in Computer Science, Machine Learning, Statistics, or related fields; or an equivalent qualification acquired through other avenues.Experience working with generative models for evaluation and/or product development, and up-to-date knowledge of common challenges and failures.Strong engineering skills and experience in writing production-quality code in Python.Deep experience in foundation model-based AI programming (i.e.: using DSPy for optimizing foundation model prompts, for example) and a drive to innovate in this space.Experience working with noisy, crowd-based data labels and human evaluations.Publication record in relevant conferences (e.g., NeurIPS, ICML, ICLR, EMNLP, etc.)Experience working on Responsible AI and AI SafetyStrong organizational and operational skills working with large, multi-functional, and diverse teams.Curiosity about fairness and bias in generative AI systems, and a strong desire to help make the technology more equitable.