What are people building in computer vision, and what's still painful? [D]

A developer with experience building computer vision models, particularly for mobile optimization like the food recognition model for MyFitnessPal, is seeking to understand current projects and pain points in the computer vision community. They are looking for problems where they can develop useful tooling and are interested in discovering forums or communities where these topics are discussed and help is sought.

This post offers a direct, first-person account from an experienced developer actively seeking to identify unmet needs and tooling opportunities within the computer vision community, unlike typical retrospective analyses.