Description

In this role, you will lead the development of cutting-edge post-training strategies for AI models, focusing on enhancing reasoning and instruction-following capabilities. You will collaborate across teams to ensure high-quality performance across different modalities.

Requirements

• •PhD in machine learning, artificial intelligence, or computer science or equivalent practical experience
• •Strong background in Large Language Models, Reinforcement Learning, or preference learning
• •Research interest in aligning AI systems with human feedback and utility
• •Familiarity with experiment design and analyzing large-scale user data
• •Strong coding and communication skills
• •Experience with RLHF or DPO
• •Experience building or improving reward models and conducting human evaluation studies
• •Proven track record of publications in top-tier conferences
• •Experience with Chain-of-Thought reasoning research or process-based supervision
• •Deep understanding and experience training models from scratch or using self-play/self-improvement techniques

Tasks

• •Design and validate novel post-training pipelines for frontier-class models
• •Lead research into next-generation Reward Models
• •Investigate new architectures for Reward Modeling
• •Reduce reward hacking in preference data
• •Improve signal-to-noise ratios in preference data
• •Develop innovative methods to enhance internal reasoning capabilities
• •Focus on correctness and logic in multi-step tasks
• •Revamp and optimize RL prompts and feedback mechanisms
• •Create robust mechanisms to convert user signals into training data
• •Collaborate across teams to apply advanced recipes to various model sizes and modalities