Company Background

Global Fishing Watch is an international, non-profit organization committed to advancing ocean governance through increased transparency. We create and publicly share knowledge about human activity at sea to enable fair and sustainable use of our ocean. Founded in 2015 through a collaboration between Oceana, SkyTruth, and Google, GFW became an independent non-profit organization in 2017. Using cutting-edge technology, we create and publicly share map visualizations, data and analysis tools to enable scientific research and drive a transformation in how we manage our ocean. By 2030, we aim to monitor and map all commercial activity at sea, including all industrial fishing vessels, small-scale fishing activity, all large non-fishing vessels, and all fixed infrastructure such as aquaculture and oil rigs. We also plan to work with intergovernmental organizations and 30 governments around the globe to promote the adoption of transparency more widely and publicly share ocean data to drive better management of marine resources.

The Position

The Research and Innovation team at Global Fishing Watch (GFW) connects data science and machine learning experts with the scientific community to produce new datasets, publish impactful research, and empower others to use our data. This team harnesses satellite technology, machine learning, and big data to shed light on some of the most pressing issues facing the ocean.

Determining the characteristics of vessels, such as size and type of vessel, and activities like fishing is a critical responsibility of the Research and Innovation team. We developed a set of convolutional neural network models that classify vessels and detect when they are fishing or doing other activities, using movement data from vessel tracking systems and identity information from public vessel registries. The Machine Learning Engineer, working closely with other members of the Research and Innovation team, will play a key role in improving the accuracy, scalability, and applicability of these models. The initial emphasis will be on models and data pipelines related to vessel identity, with a secondary focus on vessel behavior models. They will also work closely with the GFW Engineering and Product teams to ensure solutions are compatible and scalable within our cloud infrastructure. Other activities include, but are not limited to, overseeing data labelling efforts, maintaining data documentation and code repositories, and assisting the development of research projects by GFW and our partners.

The incumbent will gain experience working with researchers in the field and will interface daily with GFW’s team of data scientists and machine learning engineers. They will develop further technical skills in programming, big data, and cloud computing while working for a globally diverse, flexible, and fully distributed organization. The successful candidate will be organized and excited to help Global Fishing Watch develop new partnerships and cutting-edge research.  

Principal Duties and Responsibilities

Model development, training, and evaluation

• Enhance machine learning models for vessel characterization (as described in the Supplementary Materials of Kroodsma et al. 2018) to improve inferred vessel characteristics, better detect changes in vessel class over time, and accurately identify multi-class vessels
• Help improve our vessel ID system and produce training datasets for vessel classification and behavior models
• Collaborate with the Product team to enhance the existing set of vessel classes to improve model accuracy, as well as detect new types of vessels, particularly non-fishing classes
• Collaborate with data scientists on the team to generate unified vessel tracking datasets from different data sources such as the automatic identification system (AIS) and numerous vessel monitoring systems (VMS) using a combination of spatiotemporal and identity matching
• Evaluate existing vessel behavior models for fishing detection and identify opportunities for improvement
• Develop new behavior models to detect other vessel activities, such as deep sea mining, seismic vessel testing, or sand mining
• Provide technical support for the development, advancement, and publication of GFW datasets and scientific papers

Internal research support 

• Work with large amounts of data from various types and sources, such as vessel activity, identity, and satellite imagery to identify trends, anomalies, and insights 
• Assist with data labelling to expand and improve training datasets
• Quality assurance of key data pipelines and research tables and coordinating their effective use by internal teams and external partners 

Candidate description

Qualifications you should have

• Bachelor’s degree and at least four years of professional experience, or an equivalent combination of education and experience, in computer science, ecology, fisheries, or a related field
• Demonstrated skills and experience with Python
• Strong foundation in mathematics and statistics
• Highly organized, analytical, detail-oriented, self-motivated and with critical thinking
• Ability to work with large datasets and visualize data effectively
• Experience with version control software and collaboration tools such as Git and GitHub
• Willingness to take ownership of projects and communicate project updates
• Ability to communicate about technical topics to less technical audiences
• Written and verbal communication skills in English
• Readiness to embrace Slack, Google Suite, Jira, Notion and other collaborative tools for remote work

Also great

• Some experience with additional programming languages, such as SQL and R
• Some experience working with cloud compute platforms and virtualized environments
• Some experience with deep learning models and machine learning frameworks like TensorFlow or PyTorch
• Familiarity with containerization tools like Docker
• Familiarity with fisheries
• Experience engaging with academic researchers and the peer-review process
• An appreciation for the complexities and rewards of collaborating in a remote, global, and inclusive environment
• Written and verbal communication in Spanish, Portuguese, French, Chinese, or other languages

The successful candidate will meet most, but not necessarily all, of the criteria above. Although it is obviously helpful, we do not expect that you already have a deep knowledge of building models or our key programming languages; we do expect that you have the aptitude to develop these skills and knowledge, and that you are excited about revealing human activity across the global ocean using these tools. If you don’t think you check all the boxes, but believe you have unique skills that make you a great fit for the role, we want to hear from you!

Please refer to the job posting for further information