How Should an AI Developer in Spain Present Hugging Face Demos, GitHub Repositories, and APIs?

An AI developer should organise models, notebooks, demos, APIs, and products by what a reviewer can verify today, then label experimental, active, shipped, and archived work explicitly.

A Hugging Face Space can demonstrate behaviour, GitHub can expose implementation, and a product URL can show a usable application. The portfolio's job is to state what you built and route the visitor to the strongest surviving evidence without implying that a prototype is a production system.

Which AI projects deserve separate portfolio entries?

Create a separate entry when the work solves a distinct problem, has a clear contribution from you, and offers evidence a reviewer can understand.

A model, evaluation tool, RAG prototype, public API, or end-user application can each qualify. Avoid turning every notebook, prompt test, or fork into a project; group experiments that belong to one system.

State whether you designed the interface, data pipeline, evaluation, deployment, or model integration. Do not claim authorship of a base model or dataset you only consumed.

What is the quick way to assemble the AI portfolio with IndieShow?

The quick way is to claim an IndieShow page and add each meaningful AI system with a precise status, description, tag, logo, and public destination.

Put live tools in Shipped or Built, ongoing evaluations in Building or Working on, and superseded experiments in Archived. Reorder the page so the most relevant reproducible or usable work appears first.

Use the optional metric only for a measurement you can identify and defend. A vague accuracy claim without task, dataset, or context is weaker than a factual description of the system.

Build your IndieShow pageClaim a handle, organise the projects in the editor, and review the $15 one-year and $30 lifetime publishing options in the dashboard.

Should an AI project link to Hugging Face, GitHub, documentation, or the live app?

Link to the surface that provides the strongest accessible proof for the audience you expect.

A live app helps product reviewers, a Space helps someone test a demo, a repository helps technical reviewers, and documentation helps API consumers. Check access without your own session and remove endpoints that expose secrets or private data.

One IndieShow card has one destination, so use the description to clarify what that destination contains and choose it deliberately.

How do you describe AI results without invented performance claims?

Describe the task, your contribution, evaluation method, and only the results you can substantiate.

Do not turn a small test set into a universal quality claim or present generated examples as a benchmark. If an experiment lacks reliable evaluation, say what it demonstrates and mark it as a prototype.

The related guides explain how to separate prototypes from launches and how to arrange multiple projects on one page.

How does IndieShow keep an AI portfolio accurate as experiments change?

IndieShow lets you keep one public handle while updating project status, copy, links, metrics, logos, and order from the dashboard.

When a Space stops running, an API is retired, or a prototype becomes a product, update the same project entry instead of leaving a stale trail across bios. External platforms remain the technical proof.

IndieShow becomes the maintained index: a concise record of what is usable, what is being tested, and what you chose to archive.