Data Scientist
Description
Data Scientist
RPI is seeking qualified candidates to increase efficiency of data processing and interpretation of acoustic and optical imagery from sensors primarily embedded in uncrewed systems (autonomous underwater vehicles, remotely operated vehicles, uncrewed airborne systems). Specifically, the candidate will have expertise using and modifying, or developing new scripts in Python that will automate tasks in data processing or interpretation, data fusion from multiple sensors, signal processing, or data visualization related to georeferenced imagery from acoustic imaging, optical imaging or laser profiling sensors.
Overview
- Seeking highly qualified candidates for a Data Scientist position on an interdisciplinary research team of contract and federal employees supporting the NOAA National Centers for Coastal Ocean Science (NCCOS) Marine Spatial Ecology Division (MSE).
- NCCOS and MSE is a nationally recognized science program that conducts spatial ecological analysis, statistical modeling, ecological forecasting, and predictive mapping to support marine ecosystem management, conservation, and spatial planning. The candidate will be employed by RPI through a labor contract supporting the NOAA National Ocean Service in Beaufort, NC or Silver Spring, MD.
- We seek candidates with demonstrated expertise in signal processing and creating macros or scripts to automate processing of data from remote or underwater sensors, and artificial intelligence / machine learning for computer vision and related tasks to characterize marine environments. MSE analyzes imagery to characterize underwater habitats, from the species and benthic substrates, to the larger-scale, ecological communities and ecosystems. Automation and machine learning present scalable solutions for the MSE processing routines, specifically to 1) reduce human annotation effort, 2) reduce human-induced errors and bias, 3) enhance reproducibility for auditing, and 4) optimize pattern detection unapparent through manual analysis.
Core Responsibilities
- Assist in marine mapping and characterization related projects by working with federal investigators to develop solutions for automation at scale.
- Work with other NCCOS data scientists to facilitate deployment of automation and AI solutions on local machines or cloud environments.
- Implement and / or develop computer vision and machine learning algorithms for analyses, including algorithms for model evaluation and selection, validation, skill assessment, and ground-truthing.
- Lead and contribute to peer-reviewed publications, presentations, and technical memoranda.
- Provide analytic and technical guidance to other team members in deploying AI models and other automated tasks.
- Travel to federal and state laboratories, academic institutions, and field missions as part of collaborative research projects (<20% of time).
Required
- Master's or doctoral degree in the fields of Computer Science, Machine Learning, Information Technology, Computational Biology or Ecology, or similarly related field
- Minimum of 1-3 years of experience with developing code in applied science, commercial, or business enterprises.
- Programming skills in Python, R, Javascript, or other scripting languages
- Experience with machine learning and computer vision libraries such as PyTorch, TensorFlow, Keras, OpenCV
- Experience working on computer vision applications including but not limited to: image classification, semantic segmentation, instance segmentation, object detection, and point cloud classification.
- Familiarity with deploying code and models to Cloud providers such as Azure, and experience with Azure's Machine Learning Studio
- Experience working with open-source libraries to prepare data, fine-tune models, perform inference, and deploy models to cloud / production environments.
- Proficiency in programming and file version control tools such as Git, and GitHub is expected.
- Demonstrated excellence in written and oral scientific communication skills;
- Demonstrated experience working independently and with a team.
- Ability to work effectively in a team-oriented, multi-project, multi-disciplinary environment;
- Non-U.S. citizens must possess current documentation authorizing employment in the United States and meet the minimum security requirements for access to federal facilities;
Preferred
- Experience working with environmental, ecological, or marine datasets.
- Experience working with point cloud datasets derived from Structure-from-Motion, LiDAR, or sonar; manipulating data stored in common open-source formats. Open3D, and CloudCompare would be an advantage.
- Experience working with geospatial data (optical, side-scan, synthetic aperture sonar, underwater lasers) obtained from marine-related vehicles including ROVs, AUV, ASVs, or previous work with autonomous driving vehicles.
- Experience with GIS software including ArcGIS, QGIS, as well as geospatial coding libraries, would be beneficial
- Experience with containerization frameworks such as Docker, as well as virtual environments including Pip and Anaconda, and packaging Python projects.
- Experience in creating scripts and programs with GUIs.
- Ability to go to sea aboard a research vessel or other field research
A National Agency Check and Inquiries (NACI) background check and fingerprinting will be required.
Skills
Sign in to read and join the discussion.