Tech Stack

Description

Work with cutting-edge vision-language models and embeddings to extract insights from Planet's archive of geospatial imagery. Contribute to developing a 'Queryable Earth' by designing ML workflows, optimizing models, and collaborating with research scientists on real-world geospatial intelligence projects.

Requirements

• Pursuing graduate degree in CS, Data Science, Remote Sensing, or related field
• Proficiency in Python and deep learning frameworks (PyTorch, TensorFlow)
• Familiarity with computer vision and NLP techniques
• Knowledge of geospatial data formats and tools (GDAL, GeoPandas, Rasterio)
• Experience with vector databases and similarity search algorithms is a plus

Responsibilities

• Optimize and fine-tune novel VLMs for Earth observation data
• Create and optimize embeddings for large-scale satellite imagery datasets
• Design and execute machine learning workflows for geospatial analysis
• Contribute to 'Queryable Earth' functionality using natural language queries
• Evaluate algorithms for feature detection and classification in satellite imagery