Senior AI Engineer Self-Supervised Learning
Amazon-owned Swiss robotics company developing wheeled-legged, Physical-AI robots for autonomous doorstep delivery.
Funding history
About RIVR
RIVR develops full-stack autonomous delivery robots that combine wheels, legs, AI software, and remote operational support to navigate the last mile and final 100 yards, including curbs, stairs, gates, and doorsteps.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will build scalable data pipelines that transform raw multimodal robotic sensor data into curated datasets for neural-network training. You will develop self-supervised learning methods, pseudo-labeling workflows, evaluation benchmarks, and data access patterns, while partnering with model teams on data strategies and interfaces.
Requirements
- Master’s degree or higher in Computer Science, Machine Learning, Robotics, or a relevant field
- At least three years of industry or research experience
- Expertise in self-supervised learning and representation learning with multimodal sensor data
- Experience building large-scale machine-learning data-processing pipelines
- Understanding of robotic sensor data
- Python and PyTorch experience, including custom DataLoaders
- Experience with MLOps practices and data versioning tools
Responsibilities
- Design, build, and maintain scalable pipelines for multimodal robotic sensor data
- Develop self-supervised and representation learning algorithms
- Generate features, discover patterns, and create pseudo-labels from unlabeled data
- Define data requirements, APIs, and strategies with VLA and reinforcement learning teams
- Architect evaluation strategies, benchmarks, and datasets
- Own data integration workflows, data loaders, and access patterns
- Research data curation, active learning, and anomaly-detection techniques