Applied Data & Evaluation Engineer for World Models & Robotics

Orbifold AI; Palo Alto (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Orbifold AI in Palo Alto is seeking a Member of Technical Staff to build data and evaluation foundations for world models and embodied AI systems. This role centers on real-world system performance and scalable data pipelines, not purely theoretical work.

You will design, build, and iterate data workflows, evaluation frameworks, and tooling with robotics and world-model teams, translating failures into concrete data curation and training actions for partner deployments.

Qualifications

• Experience in building end-to-end data pipelines for ML systems.
• Familiarity with robotics, embodied AI, and multimodal data.
• Understanding of evaluation frameworks and failure mode analysis.
• Hands-on experience with production ML systems.

Responsibilities

• Co-design data pipelines and model training workflows end to end with robotics and world model teams.
• Define how multimodal data should be structured and indexed for training and evaluation.
• Build scalable systems for data ingestion, cleaning, annotation, and curation across failure modes.
• Work on training or fine-tuning models for perception, policy learning, or multimodal reasoning including automated critics.
• Iterate on datasets and training setups based on downstream model performance.
• Build comprehensive evaluation frameworks, including failure taxonomies and distribution shift.
• Bridge gaps between raw data, dataset design, and model behavior, translating evaluation findings into data and training recommendations.

Skills

Robotics

Embodied AI

World models

Multimodal reasoning

Reinforcement learning

Data pipelines

Job description

Orbifold AI in Palo Alto is seeking a Member of Technical Staff to build data and evaluation foundations for world models and embodied AI systems. This role centers on real-world system performance and scalable data pipelines, not purely theoretical work.

You will design, build, and iterate data workflows, evaluation frameworks, and tooling with robotics and world-model teams, translating failures into concrete data curation and training actions for partner deployments.