AI-powered camera systems for real-time fish farm monitoring and decision-making
Aquabyte deploys camera-based monitoring systems into fish farms, using PyTorch and edge ML (NVIDIA Jetson) to extract health and behavioral metrics from live video feeds. The tech stack—Python, Snowflake, dbt, Apache Airflow—suggests a data pipeline designed to ingest and transform high-volume visual data into actionable insights. Active projects span 3D scene reconstruction, weight estimation, feeding plan automation, and hardware observability, indicating Aquabyte is scaling from basic fish counting toward predictive farm operations.
Aquabyte builds hardware-embedded AI systems for aquaculture operators, ranging from open-pen sea farms to closed land-based facilities. The platform captures video from underwater cameras, runs inference to measure fish weight, growth, lice load, and behavioral markers (swim speed, tilt, breathing patterns), then surfaces findings through a data pipeline to farmers' dashboards. Core use cases include early disease detection, lice outbreak prevention, and feed optimization. The company operates across Norway, the US, and Chile, with an engineering-focused team concentrated on edge deployment and data infrastructure.
Camera-based visual monitoring with edge ML inference (PyTorch on NVIDIA Jetson), backed by cloud pipelines (Snowflake, dbt, Apache Airflow) for data aggregation and decision support.
Fish weight, growth rates, skin health, fin damage, sea lice counts, swim speed, swim tilt, and breathing index—used to detect health issues and optimize feeding and treatment timing.
Other companies in the same industry, closest in size