AI Alignment Engineer

Zizy Inc.; New York, Northern (NY, KY); Hybrid; USD 120,000 - 220,000; Full time

Benefits offered by this job

Mental health support

Generous time off

Parental leave and family-planning

Annual learning & development stipend

Job summary

Zizy Inc. is seeking research engineers and research scientists to design and implement experiments for alignment research, focusing on scalable training signals for AGI alignment with human intent and safety.

You will design experiments to measure scalable oversight, study generalization, manage large datasets from interpretability experiments, and develop LLM-based alignment methods in a fast-paced, collaborative team environment.

Qualifications

• Experience designing and running ML experiments.
• Interest in alignment research and AI safety.
• Ability to work in fast-paced, collaborative research environments.

Responsibilities

• Design experiments to measure scalable oversight techniques (AI-assisted feedback and debate).
• Study generalization to see if models trained on easy problems solve hard ones.
• Manage large datasets from interpretability experiments and create visualizations.
• Develop experiments to test chain-of-thought reasoning reflecting model cognition.
• Investigate how training against a reward signal affects outputs.
• Explore methods to understand/predict model behaviors, incl. anomalous circuits or catastrophic outputs.
• Design novel approaches for using LLMs in alignment research.

Skills

ML algorithms

PyTorch

Job description

The Zizy Alignment Team is working on technical approaches to ensure that artificial general intelligence (AGI) aligns with human values and follows human intent. Through scientific experimentation, we explore the scalability of alignment techniques and identify potential breaking points. Our approach to alignment research primarily focuses on engineering a scalable training signal for advanced AI systems that aligns with human intent. This approach is built upon three main pillars:

• Training AI systems using human feedback, such as learning from human instructions
• Training AI systems to assist human evaluation, such as training models to self-critique
• Training AI systems to conduct alignment research, such as using language models to explain neurons in language models

About the Role

We are seeking research engineers and research scientists to help design and implement experiments for alignment research. Responsibilities may include:

• Designing experiments to measure the effectiveness of scalable oversight techniques such as AI-assisted feedback and debate
• Studying generalization to see when AI systems trained on easy problem can solve hard problems
• Managing large datasets from interpretability experiments and creating visualizations to explore interpretability data
• Developing experiments to test how well chain of thought reasoning reflects model cognition
• Investigating situations when training against a reward signal causes model outputs to deteriorate
• Exploring methods to understand and predict model behaviors, such as finding inputs causing anomalous circuits or catastrophic outputs
• Designing novel approaches for using LLMs in alignment research

You might thrive in this role if you:

• Are excited about Zizy's mission
• Possess a strong curiosity about aligning and understanding ML models, and are motivated to use your career to address this challenge
• Enjoy fast-paced, collaborative, and cutting-edge research environments
• Have experience implementing ML algorithms (e.g., PyTorch)
• Want to ensure that powerful AI systems stay under human control

Zizy is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity. We push the boundaries of the capabilities of AI systems and seek to safely deploy them to the world through our products. AI is an extremely powerful tool that must be created with safety and human needs at its core, and to achieve our mission, we must encompass and value the many different perspectives, voices, and experiences that form the full spectrum of humanity.
At Zizy, we believe artificial intelligence has the potential to help people solve immense global challenges, and we want the upside of AI to be widely shared. Join us in shaping the future of technology.

Benefits and Perks

• Mental health and wellness support
• Generous time off policy
• Paid parental leave and family-planning support
• Annual learning & development stipend

We are an equal opportunity employer and do not discriminate on the basis of race, religion, national origin, gender, sexual orientation, age, veteran status, disability or other legally protected statuses. Pursuant to the NewYork Fair Chance Ordinance, we will consider qualified applicants with arrest and conviction records.

We are committed to providing reasonable accommodations to applicants with disabilities, and requests can be made via our careers email.