Data Scientist

Project Overview:

The goal of this project is to revolutionize echo classification by developing processing techniques and building a cloud-based framework that will promote development and implementation of advanced analytical methods to classify acoustic backscatter to the lowest taxonomic level possible in near-real time. Artificial intelligence (AI), machine learning (ML), and advanced statistical techniques (e.g., inverse and Bayesian approaches) comprise advanced analytical methods that are needed to significantly advance echo classification through the combination of high-resolution echosounder, modeled backscatter, and habitat data. Once various approaches are explored, developed, tested, and compared, a toolbox of advanced analytical tools for echo classification will be implemented in the cloud for use by researchers that process and analyze active acoustic data.

Job Responsibilities:

• Utilize validated active acoustic and co-variate datasets for inclusion in the development and validation of analytical methods for echo classification.
• Work with project team members to complete shared tasks
• Assist with the development of data transfer and sharing agreements, pipelines, and guidelines for cloud storage and computing
• Assist with populating cloud storage repositories and implementing cloud computing
• Review current methodologies applicable to echo classification across diverse datasets and identify new advanced methods for investigation.
• Propose data processing pipelines (acquisition to echo classification) for future testing that can be presented as a series of flow charts
• Identify AI/ML, Bayesian, or other modeling approaches to classify backscatter collected from multifrequency CW EK60/EK80 and FM EK80 data.
• Develop and test a set of software tools to conduct echo classification from CW, FM, and low-frequency data.
• Encode AI/ML and Bayesian approaches for local and cloud implementation.
• Test encoded methods on validated datasets and assess precision, recall, and other metrics of performance.
• Refine methods to produce stable and acceptable processing algorithms.
• Present results to project partners and draft a report or peer-reviewed manuscript of findings.
• Work with commercial (e.g. Echoview) and/or academic (e.g., Applied Physics Laboratory, WHOI) partners to develop and operationalize a toolbox for echo classification.
• Work with project, commercial, and academic partners.
• Incorporate encoded AI/ML and Bayesian approaches into existing software packages or develop new software packages for automated to semi-automated echo classification.
• Test toolboxes on validated datasets and unvalidated datasets.

Qualifications:

• Proficiency in software languages, e.g. Python, R
• Google Cloud Platform (GCP) HIGHLY DESIRED
• Proficiency in machine learning or other AI methods
• Experience with Bayesian methods
• Experience with or knowledge of common echosounder data softwares, e.g. Echoview, Echopype, PyEchoLab
• Experience with or knowledge of relational databases
• Experience with or knowledge of the principles and theories of fisheries acoustics, echosounder data analysis, and echo classification.
• Competent working with diverse teams in-person and remotely across multiple time zones
• Excellent verbal and written communication skills

Benefits: Competitive benefits package including health, dental, vision, life insurance coverage, 401(k) Plan, Training Programs, Accrued Paid Time Off (PTO) and Paid Holidays.

Equal Opportunity Employer/Veterans/Disabled