As a member of the Perception Machine Learning team, you will be responsible for proposing, prototyping and deploying Machine Learning models that address the needs of the perception and navigation systems in our robots. You will work closely with other engineers in the perception, navigation and system robotics teams to identify areas where ML solutions could help improve our product and then develop and deploy these modeling solutions on the robot. This role will include both data science (e.g. model development) and engineering (e.g. data preprocessing and model deployment) components.
Responsibilities and Duties:
- Responsible for prototyping and deploying Machine Learning Computer Vision models
- Work across teams to identify areas where ML solutions can enhance our products
- Implement engineering solutions to improve data preprocessing and model deployment
- MS/PhD in computer science, robotics or related field + related industry experience (internship counts)
- Or B.S. in computer science (with emphasis on computer vision) +2 years of related industry work
- Good theoretical understanding of Deep Learning (DL) models related to Computer Vision (CV).
- Knowledge of common image classification, object detection and tracking algorithms.
- Familiarity with recent research trends in DL/CV.
- Comfortable working with Python and/or C++.
- Experience working with one of the standard DL frameworks (TensorFlow, Keras, PyTorch).
- Experience working with common image data tools (opencv, numpy, pandas, matplotlib).
- Reasonable understanding of ML model development/deployment cycle.
- Experience working with raw (un-preprocessed) data.
- Great communication skills. This is a cross-functional role and requires ability to work across teams/domains.
Other Useful Qualifications:
- High level understanding of traditional Computer Vision algorithms
- Experience with ROS
- Experience with big data processing/streaming frameworks (Spark, Kafka, Hadoop)
- Experience deploying ML models to production
- Experience developing and optimizing ML models for running on edge hardware