Multimodal Physics ML Researcher — Causal Forecasting

Causal Labs; San Francisco (CA); On-site; USD 180,000 - 260,000; Full time

Job summary

Causal Labs in San Francisco is seeking researchers to design architectures and training recipes that turn multimodal physical observations into models that predict the future of the physical world. You will tackle sparse sensors, point clouds, and hyperspectral data at scale, pushing beyond current LLM-oriented approaches.

Join a team obsessed with rigorous experimentation, ablations, and end-to-end delivery—from data pipelines to training runs—balancing scientific curiosity with practical

Qualifications

• Strong foundations in ML with depth in at least one relevant domain.
• Experience training large-scale models and analyzing results.
• Familiarity with distributed training and scaling systems.
• Proven track record turning research into production models.

Responsibilities

• Design and implement novel model architectures for multimodal physical data.
• Solve core modeling problems for physical prediction, incl. long-horizon forecasting.
• Run experiments and ablations to link data choices to predictive performance.
• Work across ML stack from data to infrastructure to scale training runs.
• Stay up-to-date with research and bring new ideas to the project.

Skills

ML fundamentals

World models / physics-informed NNs

Large-scale model training

Distributed training

Research to production

Job description

Causal Labs in San Francisco is seeking researchers to design architectures and training recipes that turn multimodal physical observations into models that predict the future of the physical world. You will tackle sparse sensors, point clouds, and hyperspectral data at scale, pushing beyond current LLM-oriented approaches.

Join a team obsessed with rigorous experimentation, ablations, and end-to-end delivery—from data pipelines to training runs—balancing scientific curiosity with practical