Research Engineer

Goliath Partners Inc.; San Francisco (CA); On-site; USD 180,000 - 280,000; Full time

Job summary

Goliath Partners Inc. is seeking a Research Engineer to help build the data behind their next-generation world models in San Francisco. You will own multimodal datasets and run large-scale experiments, shaping model capabilities from data up.

As a Member of Technical Staff, you design datasets, build pipelines for synthetic data generation, and develop evaluations to measure improvements in frontier multimodal world models with a fast iteration loop.

Responsibilities

• Design multimodal datasets that directly shape what frontier world models learn.
• Run controlled training experiments to understand how data impacts model capabilities.
• Build large-scale pipelines for synthetic data generation, filtering, and quality control.
• Develop evaluations and benchmarks that measure meaningful model improvements.
• Work at the frontier of multimodal foundation models and world models.
• High ownership within a small, elite team operating with an exceptionally fast iteration loop.
• Directly shape the capabilities of next-generation AI systems from the data up.

Skills

Multimodal data handling

Data pipelines

Model experimentation

Synthetic data generation

Evaluation benchmarks

Job description

Goliath Partners has exclusively partnered with a fast-growing, venture-backed frontier GenAI Research Lab that is operating at serious scale & expanding.

They’re looking for a Research Engineer to help build the data behind their next-generation world models. You'll own multimodal datasets, synthetic data gen, and large-scale experimentation that directly shapes model capabilities.

As a Member of Technical Staff you will:

• Design multimodal datasets that directly shape what frontier world models learn
• Run controlled training experiments to understand how data impacts model capabilities
• Build large-scale pipelines for synthetic data generation, filtering, and quality control
• Develop evaluations and benchmarks that measure meaningful model improvements
• Work at the frontier of multimodal foundation models and world models
• High ownership within a small, elite team operating with an exceptionally fast iteration loop
• Directly shape the capabilities of next-generation AI systems from the data up