You might not realize this, but there is an enterprise version of Edge Impulse that has a ton of cool features for building (and managing) large TinyML projects. One of those features is the ability to have custom deployment targets. These are options that show up under Deployment (like all the built-in libraries and fully supported dev boards) but are controlled by you.
You can use this to automatically build your impulse into firmware for your product, let your FAEs create customized builds for customers, or to set up pipelines to output in specific library formats. Custom deployment targets are hosted in our infrastructure and are automatically exposed to all projects in your organization, so everyone in the team can use them.
Interested? We’ve just published a tutorial on this: https://docs.edgeimpulse.com/docs/building-deployment-blocks, and drop us an email at hello@edgeimpulse.com if this has peaked your interest!