Research Engineer / Research Scientist, Alignment

OpenAI; San Francisco (CA); Hybrid; USD 120,000 - 180,000; Full time

Job summary

Join a forward-thinking company as a Research Engineer/Scientist focused on AI alignment. This dynamic role involves developing innovative methodologies to ensure AI systems align with human values, particularly in complex and high-stakes scenarios. Collaborate with a passionate team to design scalable solutions, integrate human oversight, and create new evaluation methods. The position offers a hybrid work model, allowing for flexibility while contributing to groundbreaking research in AI safety and trustworthiness. If you're eager to make a meaningful impact in the field of AI, this opportunity is perfect for you.

Qualifications

• PhD or equivalent in computer science or related fields.
• Strong engineering skills in large-scale machine learning systems.

Responsibilities

• Develop and evaluate alignment capabilities for AI systems.
• Design evaluations to measure risks and alignment with human values.

Skills

Machine Learning

AI Alignment

Data Visualization

Python

TypeScript

Education

PhD in Computer Science

PhD in Data Science

PhD in Cognitive Science

Tools

PyTorch

Job description

Research Engineer / Research Scientist, Alignment

Join to apply for the Research Engineer / Research Scientist, Alignment role at OpenAI.

About The Team

The Alignment team at OpenAI is dedicated to ensuring that our AI systems are safe, trustworthy, and aligned with human values, even as they scale in complexity and capability. Our work focuses on developing methodologies for AI to follow human intent across various scenarios, including adversarial and high-stakes situations. We aim to address the most pressing challenges, ensuring our models are prepared for real-world deployment.

About The Role

As a Research Engineer / Research Scientist on the Alignment team, you will work on ensuring AI systems follow human intent in complex scenarios. Your responsibilities include designing scalable solutions for AI alignment, integrating human oversight, and developing new evaluation methods.

This role is based in San Francisco, CA, with a hybrid work model (3 days in-office) and relocation assistance available.

Responsibilities

• Develop and evaluate alignment capabilities that are subjective, context-dependent, and hard to measure.
• Design evaluations to measure risks and alignment with human values.
• Build tools to study model robustness in different situations.
• Design experiments to understand how alignment scales with compute, data, and adversarial resources.
• Develop new Human-AI interaction paradigms and oversight methods.
• Train models to be calibrated on correctness and risk.
• Innovate in AI alignment research approaches.

Qualifications

• PhD or equivalent in computer science, computational science, data science, cognitive science, or related fields.
• Strong engineering skills, especially in large-scale machine learning systems (e.g., PyTorch).
• Deep understanding of alignment algorithms and techniques.
• Experience with data visualization or data collection interfaces (e.g., TypeScript, Python).
• Enthusiasm for fast-paced, collaborative research environments.
• Interest in developing trustworthy, safe, and reliable AI models in high-stakes scenarios.

About OpenAI

OpenAI is committed to ensuring that artificial general intelligence benefits all of humanity. We push AI capabilities safely and seek diverse perspectives to shape the future of AI technology.

We are an equal opportunity employer and provide accommodations for applicants with disabilities.