Research Scientist/Engineer (AI)

TL;DR

Research Scientist/Engineer (AI): Spearheading the development of techniques to minimize hallucinations and enhance truthfulness in language models with an accent on creating robust systems that reflect true confidence levels. Focus on designing and implementing data curation pipelines, classifiers, and evaluation frameworks to ensure model accuracy and honesty.

Location: Must be based in New York or able to relocate.

Company

Anthropic is a public benefit corporation focused on creating reliable and interpretable AI systems.

What you will do

• Design and implement novel data curation pipelines for training data accuracy.
• Develop classifiers to detect hallucinations in model outputs.
• Create honesty benchmarks and evaluation frameworks.
• Implement techniques to ground model outputs in verified information.
• Design human feedback collection for correcting miscalibrated responses.
• Develop RL environments that reward truthful outputs.

Requirements

• MS/PhD in Computer Science, ML, or related field.
• Strong programming skills in Python.
• Industry experience with language model finetuning.
• Proficiency in experimental design and statistical analysis.
• Care about AI safety and accuracy.
• Experience in data science for finetuning LLMs.