LLM Research Engineer - Scientific Discovery at Cyrad Solutions in New York, New York, United States

Job description

• This role sits inside a well-resourced, interdisciplinary research organization applying frontier large language models and machine learning to molecular science and drug discovery. The environment combines foundational AI research, large-scale model development, and direct collaboration with computational and molecular scientists, with computational infrastructure built specifically for research at this scale.
• The position owns the research and engineering behind large language and multimodal models for scientific problems: architecture, pre-training, post-training, scaling, and the systems that make all of it run efficiently on high-performance compute. The mandate is not to integrate existing tools but to extend what large-scale models can do in science.
• Candidates who thrive here combine rigorous research instincts with the ability to ship real, working systems. Molecular science or drug discovery background is not required; depth and versatility in machine learning matter more.

Responsibilities

• Research, develop, and scale large language and multimodal models for complex scientific problems
• Design pre-training pipelines and distributed or parallel training systems for large models
• Explore advanced post-training methods including reinforcement learning, contrastive learning, and instruction tuning
• Build multimodal models spanning text, molecular graphs, 3D structures, time-series data, and other scientific modalities
• Optimize large-scale training and inference across high-performance computing infrastructure
• Partner with ML researchers, computational scientists, and domain experts to turn model advances into new capabilities for molecular science and drug discovery

Requirements

• Exceptional background in machine learning, computer science, mathematics, or a related quantitative field
• Deep expertise in large-scale ML systems, LLM architecture and training, and/or multimodal learning
• Strong Python programming and hands-on ML engineering ability
• Experience with distributed training, model scaling, training infrastructure, or high-performance computing
• Strong command of modern pre-training and/or post-training methods
• Demonstrated ability to conduct rigorous research and also build production-quality systems
• Exceptional record of academic, research, technical, or professional achievement
• Ability to work in the New York City office three days per week
• Exposure to molecular science, structural biology, or drug discovery problems (helpful, not required)
• Experience building models over non-text scientific modalities such as graphs, 3D structures, or time series

Skills

Required

• Machine learning
• Computer science
• Mathematics
• Large-scale ML systems
• LLM architecture and training
• Multimodal learning
• Python programming
• ML engineering
• Distributed training
• Model scaling
• Training infrastructure
• High-performance computing

Preferred

• Molecular science
• Structural biology
• Drug discovery

Benefits

• $300K to $800K base salary, plus substantial variable compensation and additional incentives. Hybrid schedule with three days per week in the New York City office.