Data Engineer, Multimodal Video & Audio Pipelines

Odyssey; Palo Alto (CA); On-site; USD 140,000 - 190,000; Full time

Job summary

Odyssey is building world-class data pipelines to curate large multimodal datasets for training causal, multimodal models. You will work across video and audio data, turning research questions into scalable data solutions.

You’ll collaborate with researchers to identify data requirements, improve captions, and enhance the training mix. A hands-on data role that blends engineering with research is ideal for those who love data at scale.

Qualifications

• Must enjoy cleaning and curating large multimodal datasets.
• Experience turning research requirements into data pipelines.
• Comfortable working with video and audio data at scale.

Responsibilities

• Build and run data pipelines for large-scale multimodal datasets (video, robotics, audio).
• Extract signals from raw data (speakers, background audio, video features).
• Help balance and curate training data.

Skills

Data engineering

Multimodal data

Video processing

Audio processing

Data pipelines

Research collaboration

Job description

Odyssey is building world-class data pipelines to curate large multimodal datasets for training causal, multimodal models. You will work across video and audio data, turning research questions into scalable data solutions.

You’ll collaborate with researchers to identify data requirements, improve captions, and enhance the training mix. A hands-on data role that blends engineering with research is ideal for those who love data at scale.