Remote Robotics Research Scientist – World Models

Innodata; Ridgefield Park (NJ); Remote; USD 160,000 - 185,000; Full time

Job summary

Innodata is seeking a Research Scientist in Robotics & World Models to own the science of robot data, partnering with customers and frontier labs to define data specifications and evaluation criteria for foundation models. The role focuses on decisions about real-world vs simulated data, and mapping capture modalities to model needs.

The candidate will run experiments, publish benchmarks, and collaborate with data capture and synthetic-data teams to enable high-quality, transferable robotics

Qualifications

• Hands-on industry experience in robot learning or robotics ML (4+ years).
• Bachelor's degree in CS/EE/Robotics; MS/PhD preferred.
• Trained and evaluated robot policies with PyTorch.
• Experience with robotics data formats and standards.
• Hands-on experience with simulation and synthetic data (NVIDIA Isaac Sim, MuJoCo, etc.).
• Experience with teleoperation or egocentric data collection.
• First-author publications or strong open-source contributions recognized.
• Ability to communicate data and modeling decisions clearly with rigorous experiments.

Responsibilities

• Define how to design, structure, and evaluate data for robot foundation models and validate choices experimentally.
• Translate robotics model requirements into concrete data specifications and evaluation criteria.
• Decide what to capture in the real world vs. simulation and curate training mixes across datasets.
• Guide data capture modalities and synthetic pipelines to align with model needs.
• Build evaluation methods predicting real-world transfer and sim-to-real fidelity.
• Run experiments to prove data decisions improve model performance.

Skills

Robot learning

PyTorch

Reinforcement learning

Simulation data

Data curation

Education

Bachelor's degree in CS/EE/Robotics

MS/PhD preferred

Tools

NVIDIA Isaac Sim

MuJoCo

HuggingFace

Open X-Embodiment

Job description

Innodata is seeking a Research Scientist in Robotics & World Models to own the science of robot data, partnering with customers and frontier labs to define data specifications and evaluation criteria for foundation models. The role focuses on decisions about real-world vs simulated data, and mapping capture modalities to model needs.

The candidate will run experiments, publish benchmarks, and collaborate with data capture and synthetic-data teams to enable high-quality, transferable robotics