Cloud to Street is a remote sensing platform that rapidly maps floods around the world and delivers risk analytics to the most vulnerable areas at a fraction of the cost and time of traditional flood modeling. Our mission is to ensure that all vulnerable communities – and the governments and companies that serve them – have the data and support they need to prepare and respond to increased catastrophic flooding. In the next 5 years, we aim to enable new flood protection and
insurance for 10 million people worldwide.

We are looking for a best-in-class machine learning and computer vision specialist to improve our flood maps. This person will lead the development and deployment of ML models. She or he will work closely with staff remote sensing scientists and the climatologist to build integrated decision-making tools from multiple sensors and data sources. This is an opportunity where we have a specific machine learning/computer vision problem with immediate implications for humanitarian aid and decision-making, and with room for creativity for finding the best solution regardless of approach (e.g., classification and regression trees, convolutional neural networks, etc.).

MORE ABOUT US AT CLOUDTOSTREET.INFO

Over the past five years and in partnership with Google Earth Outreach, the Dartmouth Flood Observatory, and others, we built proprietary tools to harness over 12 public and commercial satellites and community-generated data on the ground. Today, in partnership with the UN, the World Bank and others, we have helped disaster users prepare for and respond to flooding in some of the most impacted areas of the world including: the Indian states of Uttarakhand and Tamil Nadu,
Argentina, Senegal, Ethiopia, South Sudan, the Republic of the Congo, and more. In 2019, we will publicly release the world’s largest database of flood maps, expand to serve five to ten more countries, and reduce loss from flooding for potentially thousands of people.

SCOPE OF WORK

  • Develop machine learning and computer vision models to perform data imputation on no-data areas within Cloud to Street flood maps for different sensors due to cloud, cloud-shadow, or flash flooding
  • Develop new techniques for weighing information from multiple flood maps with differing spatial resolution and uncertainties
  • Integrate machine learning and computer vision models with existing Cloud to Street flood detection algorithms that are based in Google Earth Engine using the Python API
  • Deploy models for delivery in a production environment integrated with Cloud to Street’s existing web tools
  • Manage databases and data assets in Google Cloud Storage, Google Earth Engine Assets, and GitHub
  • Technical support for clients and assist with final written reports and training materials

KEY COMPETENCES, TECHNICAL BACKGROUND, AND EXPERIENCE REQUIRED

  • BA in computer science, applied statistics, geosciences, or related field, with focus on analysis of satellite or other geospatial data
  • Fundamental knowledge of machine learning and computer vision techniques
  • Experience with deploying machine learning and computer vision models in a production environment
  • Proven track record developing new machine learning and computer vision models
  • Commitment to justice, diversity, science, and solidarity with vulnerable communities
  • Ability to work on remote team with a scrappy startup mentality
  • Useful experience:
    • Masters degree in computer science, applied statistics, geosciences, or related field
    • Experience with Google Earth Engine Javascript and Python APIs
    • Experience using the Google Cloud Ecosystem
    • Experience using geospatial interpolation techniques
    • Experience with machine learning or computer vision for flood detection and a conceptual understanding of surface water hydrology

BENEFITS

  • Office space in Brooklyn, New York
  • Competitive compensation

HOW TO APPLY

Applicants are requested to send their submissions to hiring@cloudtostreet.info with:

  1. Subject line: Machine Learning Data Science Consultant, Cloud to Street
  2. Relevant publications or other scientific work
  3. Attached CV/Resume
  4. Paragraph expressing interest within the submission email

APPLICATION DEADLINE: Applications will be reviewed until January 30 with the intent to start the right candidate part time in mid February and full time in late April. Cloud to Street it committed to building an inclusive and diverse company. Women, people of color, and individuals with disabilities are especially encouraged to apply.