Senior Data Scientist; Geospatial AI

Nabat is a climate-tech startup based in the UAE using advanced technology to protect and restore natural ecosystems with precision and scale! We operate at the intersection of ecological & environmental sciences, GIS, AI, and robotics. Our mission is to enable data-driven, science-backed ecosystem management and restoration at the speed and scale necessary to address the critical challenges of climate change, deforestation, land erosion, and biodiversity loss.

Our comprehensive solution - consisting of ecosystem management software, precision aerial seeding drones, AI & ecological models, and restoration services - is already being used to plant millions of mangroves in the UAE. But we’re just getting started, as we build our platform to serve the needs of coastal, marine and arid ecosystems around the world.

If this mission resonates with you, let’s talk.

We’re building our product development team from the ground up and are looking for highly talented, self-driven Senior Data Scientists with proven experience using cutting-edge data science techniques to extract science-backed insights from large, complex geospatial data sets.

• Act as senior/staff-level individual contributor to own the development of analytical and predictive models for Nabat’s AI-powered ecosystem management platform, using a variety of AI/ML and statistical techniques.
• Research, develop, train, validate, refine and deploy models that use deep domain knowledge to extract user-friendly scientific and ecological insights from geospatial data. Some of the problems you might be working on using satellite and drone imagery:

• land use/cover classification
• identification of plant native and invasive species
• estimating carbon stock and biodiversity
• modeling and predicting the outcome of different restoration scenarios
• creating optimal flight paths for multi-drone seeding operations
• detecting ecosystems at risk
• measuring plant-level growth metrics
• high-throughput phenotyping, and so on.
• In close collaboration with software and data engineers, design and implement well-architected, scalable, robust, high-performance, AI/ML-powered systems for ingestion, storage, processing, analysis, insight extraction, and visualization of large amounts of geospatial time-series data (geotagged, ultra-high-resolution RGB, multi- and hyper-spectral imagery, KML files, LIDAR point clouds, etc from satellites and drones)
• Communicate and collaborate cross-functionally with product managers, software, hardware and data engineers, ecologists, GIS data analysts, drone pilots, operations folks, and end users to deliver key initiatives
• Knowledge transfer from and continuous engagement with researchers (from Technology Innovation Institute) and external partners to understand and take ownership what has been built to date, become the subject matter expert for your product domain, and lead the research & development roadmap for your domain

• A strong research/academic foundation with at least a Masters degree in engineering, data science, mathematics, ecological or environmental science, or similar fields, and ideally published research in peer-reviewed journals.
• Extensive, hands-on development experience with AI/ML model development in the geospatial, computer vision and/or remote sensing domain, including
• Familiarity with the latest machine learning techniques and models for image analysis, object detection, semantic segmentation, pattern recognition, predicting outcomes, anomaly detection, or similar using large geospatial datasets (RGB, multi- and hyper-spectral drone imagery, LIDAR point clouds, satellite imagery, etc)
• Experience combining data from multiple sources (sensor fusion) such as drones, satellites, ground truth, in field sensors or from multiple models (ensemble methods) to improve performance
• Researching appropriate models and techniques given a problem space, scientific and business context, and validating different options before recommending a solution
• Experience using supervised and unsupervised learning, deep learning, CNNs, RNNs, GANs, and the like in the geospatial domain
• Determinin…