Skills
About the Role
Motional is building technology that helps autonomous vehicles uncover critical intelligence from massive, multimodal sensor data. On the Data Mining team, you’ll help design the “brain” behind our Omnitag framework—creating large, multimodal teacher models and distilling them into highly efficient student models for fast, near real-time discovery of edge cases and long-tail scenarios.
Responsibilities
- Design and develop large-scale multimodal teacher models for world understanding
- Distill teacher knowledge into hyper-efficient student models suitable for high-throughput mining
- Advance retrieval optimization and reasoning systems that improve long-tail and error discovery
- Collaborate with cross-functional teams to translate model performance into measurable improvements
- Contribute to scalable training and inference pipelines for multimodal datasets
Requirements
- Strong background in machine learning with experience in representation learning
- Experience with retrieval optimization and systems for efficient querying or search
- Proficiency building and optimizing ML models for large-scale multimodal data
- Solid software engineering skills for production-grade ML pipelines
- Demonstrated ability to iterate on models based on real-world performance signals
Benefits
- Opportunity to work on impactful autonomous driving intelligence and data mining
- Collaborative team environment focused on research-to-production outcomes
- Work on cutting-edge multimodal learning and efficient model distillation