Multimodal Data Engineer: End-to-End AI Datasets

Abaka; Mountain View (CA); On-site; USD 120,000 - 225,000; Full time

Benefits offered by this job

Equity

Health/dental/vision coverage

PTO

Flexible schedule

Job summary

Abaka AI in the United States is hiring a Multimodal Data Engineer to build the systems and workflows that turn complex data requirements into high-quality datasets for some of the world’s most advanced AI teams.

You’ll own projects from initial requirements through final delivery: assessing feasibility, defining quality standards, building scalable pipelines, coordinating technical work, and ensuring datasets meet agreed specifications.

Qualifications

• Minimum 3+ years building large-scale data systems with ownership of production workflows.
• Experience delivering end-to-end datasets or data systems for external or internal teams.
• Hands-on with multiple data modalities (text, images, audio, video, 3D).
• Proficient with Spark, Ray, Flink and orchestration tools like Airflow, Dagster or Argo.
• Understanding of cloud storage, compute provisioning, and data privacy basics.

Responsibilities

• Own projects from scoping to delivery, breaking work into tasks with clear owners.
• Build scalable pipelines for multimodal data sourcing, processing, and delivery.
• Define quality gates and acceptance criteria before work begins.
• Collaborate with engineers, annotation teams, and external vendors.
• Monitor costs, yield, and dataset quality to improve throughput.

Skills

Data pipelines

End-to-end datasets

Multimodal data

Spark / Ray / Flink

Airflow / Dagster / Argo

Cost monitoring

Data privacy & security

Tools

Spark

Ray

Flink

Airflow

Dagster

Argo

Cloud storage

Docker

Kubernetes

Job description

Abaka AI in the United States is hiring a Multimodal Data Engineer to build the systems and workflows that turn complex data requirements into high-quality datasets for some of the world’s most advanced AI teams.

You’ll own projects from initial requirements through final delivery: assessing feasibility, defining quality standards, building scalable pipelines, coordinating technical work, and ensuring datasets meet agreed specifications.