Research Assistant, Accountable Values

About the Role:

The AI project at the Berkman Klein Center is seeking a research assistant to support our work on accountable values in large language models for the remainder of the 2023-2024 academic year. Preferred candidates have some experience with reinforcement learning and natural language processing. RA duties will involve using the OpenAI API to train a reward model using GPT4 and helping the AI project to develop capacity to rigorously study value embedding through reinforcement learning with human feedback. Prior experience with the OpenAI API is a plus, but not necessary.

Eligibility and Additional Information:

• Applicants must be actively enrolled Harvard students
• Time commitment: 10-15 hours per week
• Applicants must posess work authorization in the US
• The University’s standard rate for RA positions is $21.00/hour. No other benefits are provided.

 
To Apply: Please send your current CV or resume and a brief cover letter summarizing your interest and experience to Tom Zick (tzick@cyber.harvard.edu) and Manon Revel (mrevel@mit.edu) with the subject line: Accountable Values RA.