AI Scientist

Responsibilities

• Propose, explore, and hands-on build core algorithms for a self-improving reasoning engine
• Push the frontier of data-efficient methods for LLM reasoning
• Probe LLMs to extract hidden knowledge and synthesize fragments into reliable, complex answers
• Design core self-improvement loops that allow systems to learn autonomously from solved problems

Requirements

• Deep, first-principles understanding of LLM reasoning and failure modes
• Expertise in ML algorithm design, search, or optimization
• Knowledge of the limitations of common LLM-training methods
• Ability to design novel, data-efficient methods for task-specific reasoning strategies
• Strong publication record