Data Engineer — Egocentric & Robotics Data Pipelines

cLocation: Onsite

Employment Type: Full-time

About CNTXT

CNTXT is building the data infrastructure powering the next generation of AI and robotics. Our team develops scalable platforms that transform raw multimodal sensor data into high-quality datasets used to train frontier AI models.

We work across computer vision, robotics, distributed systems, and machine learning infrastructure, building reliable pipelines that process massive volumes of egocentric and robotics data while maintaining the highest standards of quality, reproducibility, and scalability.

About the Role

We're looking for a Data Engineer — Egocentric & Robotics Data Pipelines to build and operate the data infrastructure behind large-scale robotics and egocentric AI datasets.

Raw data arrives from devices including GoPro cameras, ZED stereo rigs, PICO and Apple Vision Pro headsets, mobile phones, and robotics platforms. You will transform raw sensor data into clean, validated, richly annotated datasets that meet strict customer specifications.

This hands-on role sits at the intersection of distributed data processing, computer vision, robotics, machine learning infrastructure, and cloud platforms.

What You'll Do

• Design, build, and maintain distributed data-processing pipelines using Ray / Ray Data.
• Build dependency-aware workflows running on Kubernetes (EKS).
• Process large-scale video and multimodal sensor datasets.
• Develop automated validation and quality-control systems for incoming vendor datasets.
• Validate codecs, resolution, frame rate, FOV, IMU synchronization, camera calibration, metadata integrity, and provenance.
• Build pipelines for video transcoding, re-encoding, container parsing, sensor synchronization, and dataset packaging.
• Productionize CV models including hand tracking, stereo triangulation, head pose estimation, VIO, and SLAM.
• Standardize datasets into LeRobot, MCAP, Foxglove, and RLDS formats.
• Build reliable ingestion and delivery endpoints.
• Manage dataset metadata, lineage, versioning, and reproducibility.
• Optimize throughput, partitioning, fault tolerance, and scalability.

Our Tech Stack

Python • Ray / Ray Data • Kubernetes (EKS) • KubeRay • Helm • Argo CD • ClearML • ffmpeg / ffprobe • MCAP / Foxglove • OpenCV • NumPy • PyTorch • AWS • PostgreSQL

What We're Looking For

• Strong Python programming skills.
• Solid software engineering fundamentals (testing, Git, code reviews, modular code).
• Experience with distributed data pipelines (Ray, Spark, Dask, Beam, or similar).
• Understanding of throughput optimization, partitioning, and fault tolerance.
• Hands-on Kubernetes experience.
• Experience with video processing, multimedia containers, sensor/IMU data, binary parsing, or automated annotation pipelines.
• Experience building scalable production data systems.

Nice to Have

• Production experience with computer vision models (hand tracking, pose estimation, depth, SLAM, VIO).
• PyTorch and ML model packaging.
• Knowledge of robotics datasets, camera calibration, 3D geometry, and stereo triangulation.
• Experience with MCAP, Foxglove, ROS, LeRobot, or RLDS.
• MLOps / GitOps experience with ClearML, Argo CD, Helm, and CI/CD.

What We Offer

• Work on cutting-edge AI and robotics infrastructure.
• Build systems powering next-generation foundation models.
• Exposure to distributed systems, computer vision, and robotics.
• Competitive compensation and career growth.