Machine Learning Researcher, Post-Training for Foundation Models

What you'll do

• Design and iterate on end-to-end post-training recipes combining SFT, Reinforcement Learning, and reasoning regimes.
• Develop and implement novel algorithms for preference optimization, model steering, and safety.
• Research methods for high-quality human and synthetic data generation and automated data filtering.
• Create curriculum learning strategies to improve model instruction following and reasoning capabilities.
• Design robust evaluation frameworks to measure model helpfulness, factuality, and real-world utility.
• Solve frontier problems in reward modeling to resist reward hacking and handle sparse rewards.
• Optimize foundation models specifically for Apple silicon and private, personal OS experiences.

What we're looking for

• Demonstrated expertise in deep learning with a focus on LLMs, post-training, or reinforcement learning.
• Strong publication record or real-world experience in deep learning and related domains.
• Proficient programming skills in Python and a deep learning framework like JAX or PyTorch.
• PhD or equivalent practical experience in Computer Science, Machine Learning, or a related technical field.
• Proven track record in post-training algorithms, techniques, and best practices for large foundation models (preferred).
• Deep experience with human data labeling, synthetic data generation, and data quality assessment for foundation models (preferred).
• Experience in improving model performance on reasoning tasks such as math, coding, and logic (preferred).
• Experience training SOTA large models at scale and familiarity with distributed training challenges (preferred).