PeopleofSuperintelligence

Researcher

Mike Lambert

Mike Lambert is a software engineer and AI researcher whose work spans recommendation systems and language-model safety. He studied EE/CS at MIT, worked as a software engineer at YouTube and Google, built DanceDeets, and worked on bikes and scooters at Lyft. In 2024, he contributed core training infrastructure to Sleeper Agents and coauthored Many-shot Jailbreaking. His foundational research informed Anthropic’s 2025 advanced tool use on the Claude Developer Platform. He is a Member of Technical Staff at Anthropic and describes himself as learning machine learning and AI with an interest in alignment.

Badges

Timeline

  1. 2000–2004

    Undergraduate at MIT

    He is an undergraduate in EE/CS at MIT from 2000 to 2004.

  2. January 2008–February 2015

    Software engineer at YouTube

    In 2008, he works as a software engineer in San Francisco at YouTube and starts taking hip-hop choreography, breaking and salsa classes. In 2010, he is one of 11 authors of The YouTube video recommendation system, presented at the Fourth ACM conference on Recommender systems.

  3. 2010–2018

    Building DanceDeets

    In 2010, he moves to New York City and launches DanceDeets, a website for street-dance events. From 2015, he takes the project more seriously and releases mobile apps; by March 2017, it lists over 150,000 events. He creates and operates the platform as CEO/developer from 2011 to 2018.

  4. 2015–2018

    Software engineer at Google

    From 2015 to 2018, he works as a software engineer at Google.

  5. June 2018–June 2022

    Bikes and scooters at Lyft

    From June 2018 to June 2022, he is a principal software engineer at Lyft and works on transit bikes and scooters.

  6. January 2024

    Core infrastructure for Sleeper Agents

    In January 2024, he is a core research contributor to Sleeper Agents, an Anthropic study showing that backdoor behavior can persist through standard safety training. He helps develop the project’s core infrastructure to train the backdoored models.

  7. 2024

    Many-shot Jailbreaking at NeurIPS

    In 2024, he coauthors Many-shot Jailbreaking, published in Advances in Neural Information Processing Systems 37. The study examines long-context attacks that prompt language models with hundreds of demonstrations of undesirable behavior.

  8. As of October 2026

    Member of Technical Staff at Anthropic

    He is a Member of Technical Staff at Anthropic. In 2025, his foundational research informs advanced tool use on the Claude Developer Platform. He describes himself as learning machine learning and AI with an interest in alignment.

Career

  1. 2011–2018DanceDeetsCEO/developer
  2. AnthropicCore research contributor

Education

  1. 2004Massachusetts Institute of Technology

From the biography.

  1. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  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. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

Sources and credits

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

  1. Mike Lambert | alphaXiv alphaxiv.org. Checked 21 October 2026.
  2. The DanceDeets Story - by Mike Lambert mlambert.substack.com. Checked 21 October 2026.
  3. The YouTube video recommendation system research.google. Checked 21 October 2026.
  4. mikelambert (Mike Lambert) · GitHub github.com. Checked 21 October 2026.
  5. [2401.05566] Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training arxiv.org. Checked 21 October 2026.
  6. Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training (HTML full text) arxiv.org. Checked 21 October 2026.
  7. Many-shot Jailbreaking proceedings.neurips.cc. Checked 21 October 2026.
  8. Individual Signatories - Call for red lines to prevent unacceptable AI risks red-lines.ai. Checked 21 October 2026.
  9. Introducing advanced tool use on the Claude Developer Platform anthropic.com. Checked 21 October 2026.