Applied AI systems for agriculture.
We design and deploy locally-executable, multimodal pipelines for data-driven decision-making in rural agricultural environments.

• Field-ready ML systems: Integrating satellite imagery, multispectral drone video, sensor data, and agronomic documents.
• Offline-first architecture: All models and pipelines are designed to operate on-premise in low-connectivity regions.
• Multimodal & geospatial AI: Combining computer vision, retrieval-augmented generation (RAG), and reinforcement learning to support crop decisions and optimize supply chains.

PyTorch · TorchAO · ExecuTorch · GDAL · rasterio · scikit-learn · MLflow · GeoPandas · FAISS · LangChain

All our work is validated under real agricultural conditions in southern Chile (Ñuble, Biobío, La Araucanía, Los Lagos), supporting crops like cherries, maize, and potatoes.

Some components of our geospatial and ingestion stack are open-source to support the agri-tech and applied AI communities.

We bring frontier AI where it’s needed most: offline, explainable, and field-resilient.

🔗 LinkedIn · 🌐 australmetrics.cl · 📫 [email protected]