About the job
About Our Team
The Safety Systems team plays a critical role in ensuring the secure deployment of our advanced AI models. Our mission is to promote the responsible use of artificial intelligence, fostering a culture of transparency and trust as we work towards the safe development of AGI. We are dedicated to understanding deployment challenges and maximizing the societal benefits of AI while ensuring its safe utilization.
About the Position
In the position of Lead Researcher for Cybersecurity, you will spearhead the design, implementation, and management of a comprehensive mitigation framework to counteract severe cybersecurity threats across OpenAI's products. This pivotal role requires deep technical expertise, strong leadership capabilities, and the ability to influence cross-functional teams to create effective and scalable safeguards. You will establish technical strategies, drive implementation efforts, and ensure the integrity of our products against potential misuse.
In This Role, You Will:
- Develop and lead a comprehensive mitigation strategy addressing cybersecurity risks associated with model deployment, encompassing prevention, monitoring, detection, and enforcement.
- Ensure consistent and low-latency safeguards are integrated across products, scaling efficiently with usage and emerging model features.
- Make informed technical decisions within the cybersecurity landscape, weighing factors such as coverage, latency, utility, and user privacy.
- Collaborate with risk and threat modeling teams to align mitigation strategies with potential attacker behaviors and high-impact scenarios.
- Conduct rigorous testing, including red-teaming and stress-testing, to evaluate the effectiveness of the mitigation stack against evolving threats and across product surfaces.
You May Be A Great Fit If You:
- Are passionate about AI safety and eager to contribute to the secure application of cutting-edge AI technologies.
- Possess proven expertise in deep learning and transformer architectures.
- Are well-versed in frameworks such as PyTorch or TensorFlow.
- Have a solid understanding of data structures, algorithms, and software engineering best practices.
- Are knowledgeable about training and fine-tuning large language models, including diverse safety mechanisms.

