Autonomous Driving Software Engineer – Datasets & MLOps

About the role

ADASTEC is seeking an Autonomous Driving Software Engineer to own datasets and MLOps for its perception stack. You will collaborate with perception engineers, annotators, and model owners to transform multi‑sensor driving data into reliable, versioned training sets ready for automated labeling and training.

Key responsibilities

• Prepare, organize, and maintain multi‑modal datasets (camera, LiDAR, radar).
• Build and operate dataset versioning, storage, and access for training jobs.
• Define annotation requirements and quality checks with perception and annotation teams.
• Develop auto‑labeling workflows and tools that validate dataset and label quality.
• Monitor dataset coverage and domain shift, collaborating with model owners when data issues arise.
• Support automated training pipelines with stable, reproducible data inputs.

Required profile

• Bachelor’s or master’s in Computer Science, AI, Electrical/Electronics Engineering, Robotics, or related field.
• Minimum 2 years of professional experience as a Machine Learning Engineer.
• Excellent command of English, both written and spoken.

Required skills

• Strong Python programming skills.
• Experience with Linux, Git, Docker.
• Knowledge of object storage solutions (AWS S3 or compatible).
• Familiarity with annotation workflows and quality assurance.
• Understanding of cameras, LiDAR, or radar sensors.
• Experience with dataset versioning tools (e.g., lakeFS, DVC).
• Exposure to annotation platforms (CVAT, Labelbox, Supervisely, Segments.ai).
• Basic knowledge of ROS/ROS2 and PyTorch.