美国 AI 独角兽公司（排名前五），在新加坡设置研发中心，寻找 LLM Pre-Training Researcher （ base 新加坡），对标硅谷薪资水平，顶尖技术团队。
About the Role
Join a leading AI research company at the forefront of large language model development. As an LLM Pre-Training Researcher, you will shape the future of foundation models by working across the entire model lifecycle — from large-scale pre-training to post-training alignment. This role offers the rare opportunity to operate at the cutting edge of scaling laws, reasoning, and alignment, directly influencing how next-generation language models learn and behave in real-world applications.
Key Responsibilities
• Architect and scale large autoregressive language models, designing improved pre-training objectives to enhance reasoning and knowledge retention
• Develop mid-training strategies including continued pre-training, domain adaptation, curriculum learning, and synthetic data integration
• Advance post-training techniques such as instruction tuning, preference optimization, reinforcement learning, and inference-time compute scaling
• Curate and construct massive, high-quality text corpora for pre-training while designing synthetic data pipelines for reasoning and structured problem solving
• Train frontier-scale language models across large GPU clusters, optimizing distributed training and memory efficiency
• Build infrastructure for large-scale experimentation, ablations, and reproducibility to support scalable deployment
• Define evaluation frameworks for language intelligence including multi-step reasoning, coding, knowledge grounding, and agentic behavior
• Track capability development across training phases and close the loop between evaluation signals and model improvements
Requirements
• Strong foundation in machine learning and large language models with deep understanding of autoregressive transformers
• Hands-on experience with PyTorch and distributed training at scale in both research and production environments
• Experience with pre-training large models and post-training techniques such as instruction tuning, RLHF, or preference optimization
• Experience training frontier-scale language models from scratch (is a bonus)
• Research contributions in scaling laws, reasoning, alignment, or inference-time compute (is a bonus)
• Expertise in long-context modeling or structured reasoning systems (is a bonus)
有意者请联系 WX：sophia_liu611
或者 email： sophia.liu@dadaconsultants.com