Job Number: 22701
Workplace Type: Hybrid
Compensation: $200k plus equity
Summary of Responsibilities:
- Develop computer vision algorithms for object detection, tracking, semantic segmentation, and classification.
- Build and train deep learning models to enable complex urban scene perception and real-time analysis.
- Participate in end-to-end development: from problem statement, data aggregation, and annotation, through model design, experiments, and training, to the deployment of the optimized model on embedded platforms and iterative improvement automation.
- Automation of improvement cycles of DL models.
Summary of Qualifications:
- BS, MS, or Ph.D. in Robotics, Machine Learning, Computer Science, Electrical Engineering, or a related field.
- Must have 8+ years of Expertise in deploying real-world applied computer vision (including deep learning models) on edge devices.
- Strong Python programming and software design skills, knowledge of C++.
- Familiarity with standard tools and libraries, e.g. Pytorch, OpenCV, Tensorflow, MLflow.
- Proven track record - significant industry experience and/or publications at venues such as ICRA, RSS, IROS, or CVPR.
- Experience in automated data annotation.
- Multi-task models training.
- Semi-supervised DL models training on video data.
- Experience in design of multi-modal DL models with temporal context and geometrical constraints.
- Understanding of optimization of DL models and deployment on embedded platforms such as the Nvidia Jetson.
- Experience in CUDA programming, low-level edge model optimization using e.g. TensorRT and similar tools.
- Experience in designing automated machine learning pipelines.
We look forward to reviewing your application. We encourage everyone to apply - even if every box isn’t checked for what you are looking for or what is required.
PDSINC, LLC is an Equal Opportunity Employer.