Roadmap

We’ve integrated our AI workflows with Argo so you can run everything from start to finish with a single click.

Scaling up with Kubernetes

To handle huge datasets and models, we’re using Kubernetes to scale our compute power—whether it’s on-site or in the cloud. We’re working on containerizing our apps with Docker and using Kubernetes to manage the heavy lifting. This will allow users to uploade their media and click a single button to run a model. Below is a video of a workflow in action that runs a detection, classification and tracking model pipeline.

Object Segmentation We're also adding segmentation algorithms for projects that need to calculate the exact area of an object. This is particulalry useful for the UAV project kelp coverage studies.

Better Algorithms We're adding more advanced algorithms, like semi-supervised and self-supervised learning. These methods are great for boosting model performance, especially when dealing with rare classes. You can check out this MBARI intern paper to see how we’ve been testing this on AUV datasets.

last updated: 2026-02-09