Research Scientist, Paradigms of Intelligence

• Carry out sustained exploratory research. Apply the results of your research. Implement and share research findings with researchers and engineers.
• Review literature, identify key questions, design experiments, and interpret results.
• Collaborate in person and remotely; maintain a respectful work environment.
• Share ideas verbally and in writing; publish and present work at journals or scientific conferences.

Minimum qualifications:

• PhD degree in Computer Science, a related field, or equivalent practical experience.
• Recent and repeated evidence of independent research, shown through two or more first-authored scientific publications at the main track of the machine learning conferences (NeurIPS, ICML, or ICLR), or of the natural language processing conferences (ACL, NAACL, or EMNLP), published since 2021.
• Recent and repeated evidence of experience in large language model research (LLMs, agents, foundation models, LLM training or fine-tuning, transformers, or strongly related), shown through two or more scientific publications, published since 2021. The same papers can satisfy more than one minimum qualification.

Preferred qualifications:

• Post-doctoral experience.
• Experience in the use of LLMs in fields of program synthesis, automated code discovery, AutoML, or automated science.
• Experience in training and fine-tuning LLMs.
• Experience in programming with C++.
• Excellent Python programming expertise, including machine learning Python libraries such as JAX, TensorFlow or PyTorch.

• PhD degree in Computer Science, a related field, or equivalent practical experience.
• Recent and repeated evidence of independent research, shown through two or more first-authored scientific publications at the main track of the machine learning conferences (NeurIPS, ICML, or ICLR), or of the natural language processing conferences (ACL, NAACL, or EMNLP), published since 2021.
• Recent and repeated evidence of experience in large language model research (LLMs, agents, foundation models, LLM training or fine-tuning, transformers, or strongly related), shown through two or more scientific publications, published since 2021. The same papers can satisfy more than one minimum qualification.