Analyzes aerial imagery to find actionable information.
Computer Vision Scientist - Public Sector
Washington, DC, US
Job Description / Skills Required
In order to take our government business to the next level we are looking to hire a cleared Computer Vision Scientist. This Computer Vision Scientist will work on developing, adapting, and extending our commercial software products to fit the needs of our government customers. This role will involve developing new CV models on new (possibly classified) data sources, improving existing CV models to reach higher and higher precision and recall metrics, and deploying these models in multiple cloud environments through collaboration with other engineers. This role will be heavily focused on enhancing our platform, GO, with new and improved AI-powered capabilities.
In order to see detail and work at scale, we use computer vision and machine learning. Using convolutional neural networks, we have counted over 3 billion cars across global roads and parking lots, analyzing commercial patterns as well as city growth and development. We have recently expanded our capabilities to include detecting trucks, identifying almost 50 classes of airplanes, different types of armored vehicles, and changes in land use. Other projects include site monitoring near facilities or bases, tracking rates of deforestation, measuring poverty levels around the globe, monitoring construction rates of new roads and buildings, and predicting crop yields. This is still the beginning of our mission to understand the Earth.
This position will be based in our Arlington, Virginia office but may require some work at customer sites in the DC area. Currently, Orbital Insight is fully remote (working from home) until the end of the year, due to COVID. Occasional travel to corporate headquarters in Palo Alto, California may be required (only applicable post COVID).
This position requires active TS clearance.
Orbital Insight is a fast growing provider of geospatial data and geo-analytic services to commercial, nonprofit and public sector clients across the globe. Leveraging advancements in computer vision and cloud computing, Orbital Insight is turning millions of images into a big picture understanding of the world that is quantitatively grounded in observation, creating unprecedented transparency, and empowering global decision makers with new sources of objective, quantified socio-economic insights.
We are backed by marquee investors such as Google Ventures, Sequoia Capital, In-Q- Tel and Bloomberg Beta, Orbital Insight is rapidly expanding its commercial and public sector capabilities. Come join us if you are unafraid to try new approaches and relentlessly seek to push the state of the art, taking projects all the way from prototype to production-ready.
- Work as part of our core CV team, but focusing on public sector projects
- Design and implement highly scalable computer vision algorithms
- Rapidly iterate algorithms to continuously improve performance under a variety of conditions
- Collaborate with end-user and product owners to understand computer vision needs
- Design and implement methodologies for testing and accuracy assessment
- Support build deployment, management, and monitoring systems
- Provide production support for applications and infrastructure
- Report on deliverables to corporate leadership, our partners, and our customers
- Active TS Clearance required (must be willing to get TS/SCI with CI Poly)
- Must have designed, trained, and deployed CV models in production (industry experience required)
- MS or PhD Degree in computer science or related discipline
- 5+ years experience in computer vision, machine learning, and/or image processing
- Solid software engineering skills (python is a must)
- Proven expertise in deep learning, including semantic segmentation, object detection, and image classification with 2+ years of experience of building such models
- Experience with automated analysis of satellite or aerial imagery is a must (multispectral imagery is a plus)
- Experience with cloud computing (AWS or similar) is a plus