Staff Applied AI Engineer

Planet

3d•Remote

About The Position

We are seeking a talented Staff Artificial Intelligence (AI) engineer to join our AI Research team and contribute to our vision of creating a “Queryable Earth”. In this role, you will work with cutting-edge vision-language models (VLMs) and contrastive vision-language embedding models to extract valuable insights from our extensive archive of geospatial imagery. This is a fantastic opportunity to contribute to accelerate the shift in how geospatial data is consumed by end users. You will work with highly skilled Planeteers working on multi-disciplinary fields including satellites, space operations, image processing, data pipeline and analytics teams and work to co-develop AI/ML solutions for Planet’s geospatial imagery. Join us in revolutionizing how we understand and interact with our Planetary Dataset through the power of AI. This is a full-time, remote role, with a requirement to travel to San Francisco, HQ once a month.

Requirements

Advanced degree in Computer Science, Artificial Intelligence, Remote Sensing, or similar
12+ years expertise (or demonstrably equivalent) in Computer Science, Artificial Intelligence, Remote Sensing, or a related field
Experience with remote sensing, satellite image analysis, and geospatial data
Experience with rapid prototyping of AI Applications, especially search, LLMs, and agents, e.g. Google ADK, Model Context Protocol, CrewAI, Langchain, etc.
Extensive experience in developing and deploying AI/ML models, with a focus on geospatial applications and foundation models, embeddings, and frontier VLLMs
Excellent understanding of generative AI techniques, including LLMs and embeddings.
Proficient in Python and deep learning frameworks and high-performance distributed computing and IO frameworks using the python ecosystem, e.g. xarrays, dask, numpy, BigQuery, etc.
Expertise with computer vision and natural language processing techniques and familiarity with joint multimodal embeddings generators like CLIP and its more recent variants, as well as the operation and use of MMVLMs (multi-model vision-language models)
Familiarity with multi-dimensional geometry, statistics, linear algebra, optimization, and the internals of standard deep learning architectures
Fluency in full stack-development development and effective GUI implementation for web applications which rely on back-end scientific and AI systems
Knowledge of geospatial data formats and analysis tools (e.g., GDAL, GeoPandas, Rasterio)
Excellent problem-solving skills and ability to work in a dynamic research environment
Experience with cloud computing platforms (e.g., AWS, GCP, Azure) and big data workflows
Excellent communication and collaboration skills

Nice To Haves

Product and/or business development experience
Experience with distributed computing (Docker/K8s, Dask, Ray Serve)

Responsibilities

Develop and optimize multimodal LLM applications
Build embedding & similarity search pipelines at planetary scale—batch over xarray/Zarr cubes, index with BigQuery or PostGIS, etc.
Fine‑tune multimodal foundation models (e.g. CLIP-like models, MAEs, ViTs) for Earth‑observation tasks: change detection, land‑cover, semantic search, object counting
Design and execute machine learning workflows for geospatial analysis
Define success criteria and model benchmarks, adding instrumentation and model versioning where appropriate
Co‑design tool schemas & guardrails with backend engineers so LLM‑generated JSON plans execute safely
Collaborate with research scientists and engineers to design innovative models for remote sensing applications
Assist in automating the preprocessing and labeling geospatial data for AI tasks
Evaluate and improve algorithms for feature detection and classification in satellite imagery
Publish findings internally & externally (e.g. IGARSS, CVPR, etc.)