About the Role
Develop and optimize machine learning solutions for 3D perception tasks critical to robotic performance in unpredictable settings.
Responsibilities
- Design and implement deep learning models for 3D object detection and segmentation
- Optimize neural networks for low-latency inference on embedded hardware
- Train models using large-scale, real-world sensor data
- Collaborate with robotics engineers to integrate perception systems
- Evaluate model performance under diverse environmental conditions
- Improve data pipelines for efficient training and validation
- Debug and resolve issues in field-deployed perception systems
- Refine labeling guidelines and oversee annotation quality
- Develop tools for automated data curation and model testing
- Support deployment of models across multiple robotic platforms
- Monitor system performance and implement updates
- Work with sensor teams to enhance camera and LiDAR data quality
- Contribute to version control and model tracking systems
- Participate in cross-functional planning and technical reviews
- Ensure robustness of perception outputs in challenging lighting and weather
- Research novel approaches to improve detection accuracy
- Maintain documentation for models and training processes
- Assist in scaling training infrastructure
- Optimize model size and computational efficiency
- Collaborate on failure mode analysis and recovery strategies
- Integrate feedback from field operations into model updates
- Support safety validation for autonomous behaviors
- Work on domain adaptation techniques for generalization
- Contribute to simulation-based testing environments
- Ensure compliance with system-level timing constraints
Nice to Have
- Master’s or PhD in computer vision, machine learning, or robotics
- Published research in relevant conferences or journals
- Experience with autonomous vehicles or industrial robotics
- Knowledge of transformer architectures for perception
- Experience with real-time system constraints
- Familiarity with CUDA and GPU optimization
- Background in statistical modeling and uncertainty estimation
- Experience with large-scale training clusters
- Proficiency with Docker and containerized deployment
- Involvement in open-source ML projects
Compensation
Competitive salary with equity and benefits package
Work Arrangement
Hybrid work model with partial remote flexibility
Team
Collaborative team focused on advanced robotics and real-world AI deployment
What You’ll Do
- Build and refine machine learning models that interpret 3D sensor data for robotic decision-making
- Work on real-world challenges like detecting metal parts in variable lighting and cluttered environments
- Iterate on models using data from deployed robots in customer facilities
- Collaborate with hardware and controls teams to close the loop on perception-driven actions
Our Environment
- Fast-paced startup setting with direct impact on product development
- Access to production robots and real operational data
- Culture emphasizing technical rigor and practical problem-solving
- Opportunities to see your models operate in real industrial environments
Available for qualified candidates
