Member of Technical Staff

About Locamage

Locamage is building a universal geospatial foundation model. Most Earth AI is trained for one satellite and breaks on the next. We built ours from scratch to read across any sensor (optical, SAR, hyperspectral, thermal, elevation), any resolution, and any revisit frequency. You ask plain-text questions about what is happening on the ground and the model answers. We turn imagery archives into something searchable and actionable via embeddings.

We came out of AI research at UC Berkeley and are backed by a16z Alpha. Our customers span insurance, security, agriculture, and climate research.

The Role

As a Member of Technical Staff, you will help design, pretrain, and scale frontier geospatial models on multi-modal, planetary-scale data. We are running large-scale pretraining from scratch across thousands of XPU chips (TPUs and NVIDIA GPUs), iterating on architectures that need to handle hundreds of spectral bands, variable resolutions, and temporal sequences all at once.

You will work directly with our CEO, interface with major geospatial industry partners, and have more compute than you know what to do with. We are heavily provisioned and you will never be waiting on hardware.

What You Will Do

• Model Development: Design and pretrain large-scale geospatial foundation models that handle diverse spatial resolutions, sensor types, and temporal signals in a single architecture.
• Pretraining at Scale: Build and optimize distributed training pipelines across massive XPU clusters. Debug throughput, fix bottlenecks, push utilization.
• Dataset Curation: Wrangle petabyte-scale satellite archives into clean, efficient training pipelines. Multiple sensors, multiple providers, multiple formats.
• API and Productization: Work with the team to ship model endpoints that enterprise customers can query.

What We Are Looking For

• Background: MS or PhD student/graduate in CS, data science, aerospace/remote sensing, or a quantitative field with a focus on machine learning.
• Pretraining Experience: Pretraining experience on a hundred-thousand XPU scale. Multimodal experience is a strong plus.
• Framework Depth: Strong in JAX (preferred) or PyTorch.

Compensation and Path Forward

We offer an attractive cash package for the initial sprint, scaled to your experience and pretraining depth. For standout performers who integrate into the team and demonstrate impact, this role carries a direct path to a permanent position on the founding team with significant early-stage equity. We are based in San Francisco, but support remote work for candidates who can collaborate within US time zones.