research
Posted Jan 27, 2025Researcher, Trustworthy AI
at openai
San Francisco, United StatesRemote
Requirements
- About the team The Safety Systems team https://openai.com/safety/safety-systems is responsible for various safety work to ensure our best models can be safely deployed to the real world to benefit the society and is at the forefront of OpenAI's mission to build and deploy safe AGI, driving our commitment to AI safety and fostering a culture of trust and transparency.
- The Trustworthy AI team works on action relevant or decision relevant research to ensure we shape A(G)I keeping societal impacts in mind.
- This includes work on full stack policy problems such as building methods for public inputs into model values and understanding impacts of anthropomorphism of AI.
- Our team also works on external assurances for AI with an aim for increasing independent checks and forming additional layers of validation.
- experience (industry or similar academic experience) and proficiency in Python or similar languages - Thrive in environments involving large-scale AI systems and multimodal datasets - Enjoy working on large-scale, difficult, and nebulous problems in a well-resourced environment - Exhibit proficiency in the field of AI safety, focusing on topics like RLHF, adversarial training, robustness, LLM evaluations - Have past
- experience in interdisciplinary research - Show enthusiasm for socio-technical topics About OpenAI OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence
Experience
- - Possess 3+ years of research
Additional details
- We aim to translate nebulous policy problems to be technically tractable and measurable.
- We then use this work to inform and build interventions that increase societal readiness for increasingly intelligent systems.
- About the role We are looking to hire exceptional research scientists/engineers that can push the rigor of work needed to increase societal readiness for AGI.
- Specifically, we are looking for those that will enable us to translate nebulous policy problems to be technically tractable and measurable.
- In this role, you will enable: - Set research and strategies to study societal impacts of our models in an action-relevant manner and figure out how to tie this back into model design - Build creative methods and run experiments that enable public input into model values - Increasing rigor of external assurances by turning external findings into robust evaluations - Facilitating and growing our ability to effectively de-risk flagship model deployments in a timely manner You might thrive in this role if