PeopleofSuperintelligence

Lab leader

Nicholas Joseph

Also known as Nick Joseph

Nicholas Joseph is an Anthropic pre-training leader whose AI work spans assistant alignment, model self-evaluation and large language-model training. He coauthored studies on helpful and harmless assistants, Constitutional AI, and whether models can judge their own answers. Before 2021, he worked on computer vision at Vicarious and joined OpenAI's safety team to study code models; by May 2024, he led more than 40 people training Anthropic's large language models, including Claude. In May 2026, Andrej Karpathy joined Joseph's pre-training team with plans to use Claude to accelerate pre-training research.

Works
45

Publication numbers from OpenAlex, as of 22 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

  1. Before 2021

    Computer vision at Vicarious

    Before joining Anthropic, he works on computer vision at Vicarious and learns to do machine-learning research.

  2. Before 2021

    Code models at OpenAI

    Before joining Anthropic, he joins OpenAI's safety team and spends eight months working on code models and understanding how they would progress.

  3. December 2021

    A general assistant for alignment

    In December 2021, he coauthors A General Language Assistant as a Laboratory for Alignment, which studies prompting and ranked preference modeling for an assistant designed to be helpful, honest and harmless.

  4. April 2022

    Training helpful assistants

    He coauthors Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback in April 2022, applying preference modeling and reinforcement learning from human feedback to fine-tune helpful, harmless assistants.

  5. July 2022

    Models assess their own answers

    In July 2022, he coauthors Language Models (Mostly) Know What They Know, which tests whether models can assess their own claims and predict which questions they can answer correctly.

  6. December 2022

    Constitutional AI from feedback

    Constitutional AI: Harmlessness from AI Feedback appears in December 2022 with Joseph as a coauthor; the method trains a harmless assistant through self-improvement using principles and AI feedback.

  7. May 2024

    Leading Anthropic's training team

    By May 2024, he leads Anthropic's model-training group, managing more than 40 people focused on its large language models, including Claude.

  8. May 2026

    Pre-training at Anthropic

    Under his leadership, Andrej Karpathy joins Anthropic's pre-training team in May 2026 with plans to build a group that uses Claude to accelerate pre-training research.

Videos

YouTube videos in which Nicholas Joseph speaks or appears, newest first. A model chose them from YouTube search results. List updated 20 October 2026.

  1. Are Anthropic's AI safety policies up to the task? | Nick Joseph80,000 Hours · 22 August 2024 · 2 h 31 min
  2. WHY AND HOW OF SCALING LARGE LANGUAGE MODELS | NICHOLAS JOSEPHPyTorch · 4 January 2022 · 10 min

Video data from YouTube API Services. By watching you agree to the YouTube Terms of Service. How we use YouTube data.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Nicholas Joseph. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  2. 16% of AI career moves in the 2020s went from a university to a companyIncludes Nicholas Joseph. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  3. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Anthropic. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.
  4. Among employers with 30+ full-time cases, OpenAI leads primary technical role H-1B pay at $310,000, and Amazon, Google and Microsoft file 57% of all casesIncludes OpenAI. Across 440,669 certified US H-1B filings from 44 employers (FY2020–FY2026), OpenAI leads primary technical role base pay at $310,000 and xAI leads overall base pay at $290,000.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Nicholas Joseph yet. Send us a free-licensed photo.

  1. #197 – Nick Joseph on whether Anthropic’s AI safety policy is up to the task 80000hours.org. Checked 21 October 2026.
  2. A General Language Assistant as a Laboratory for Alignment arxiv.org. Checked 21 October 2026.
  3. A General Language Assistant as a Laboratory for Alignment arxiv.org. Checked 21 October 2026.
  4. Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback arxiv.org. Checked 21 October 2026.
  5. Language Models (Mostly) Know What They Know arxiv.org. Checked 21 October 2026.
  6. Constitutional AI: Harmlessness from AI Feedback arxiv.org. Checked 21 October 2026.
  7. Constitutional AI: Harmlessness from AI Feedback arxiv.org. Checked 21 October 2026.
  8. OpenAI co-founder Andrej Karpathy joins Anthropic’s pre-training team techcrunch.com. Checked 21 October 2026.
  9. Karpathy, Who Called Today's AI Agents 'Slop,' Joins Anthropic to Use Claude to Build the Next Claude techtimes.com. Checked 21 October 2026.