Contributing Guidelines

VectorHub is a learning hub that lives on its contributors. We are always looking for people interested in sharing their research and helping others in the community, especially as we grow. You can contribute in many ways, either by creating new content or by letting us know if content needs updating.

VectorHub (VH)'s value is in its positioning as a neutral platform for high-quality educational content.

We have seen significant growth since VH's launch in late November 2023, now hitting over 30k views (Mar ‘24) monthly across VH assets. The audience is technical, >50% ML Engineer/Data Scientist, 7% CTO level, from startups to major infrastructure / ML platform executives.

VectorHub's content is organized into three major areas:

1. Building Blocks: These cover the broad field of vector creation and retrieval. We take a step by step approach to creating a vector stack: Data Sources -> Vector Compute -> Vector Search & Management.

2. Articles: This is where contributors can share examples research and solutions to problems they have encountered while working on Information Retrieval problems.

3. Toolkit (coming soon): These are interesting apps, links, videos, tips, & tricks that aid in vector creation and retrieval.

This loom explains how to set up your contributing workflow.

When you are ready to submit:

1. Fork the VectorHub repo
2. Push all commits to your fork in the appropriate section for your content
3. Open a PR to merge content from their fork to the remote repo (superlinked/vectorhub)

Please read the style guide for guidance on how to structure and format your content.

However, some additional dos and don’ts for contributors are summarised below:

✅ Share a bullet point outline of your planned content first - To prevent a lot of work going into content that is not suitable for the platform, please share an outline, in the form of bullet points, of the content you plan to create.
✅ Link it to a use case - by making content practical in this way, people can then see how they might use your ideas in a production system.
✅ Include code for implementation - the audience is technical with a lot of builders; therefore, we always want a practical outcome people can try themselves when they read the content.
✅ Share unique insights & research - the best content on VH showcases new ways of doing things to tackle known challenges in the industry, for example, improving retrieval quality.
✅ Use simple tools/libraries or mainstream tools to demonstrate your use case in code - we appreciate that you may have preferred products when building in the real world. However, for VectorHub content, we ask that you focus on demoing how to build from first principles, using simple tools (sklearn, PyTorch, numpy and similar), no tools at all, or, if needed, mainstream libraries (i.e. llamaindex or langchain), and - to maintain the platform's neutral integrity - highlight that alternative options are available.
✅ Be realistic with submission timelines - It’s better to plan for a longer timeline and deliver early than to keep pushing back. To effectively support your submission, we need time to plan the content pipeline and promotion.
✅ Share step-by-step implementations with context - For inspiration, Andrei Kaparthy’s tutorials (here & here) on building out neural networks and GPTs from scratch are a great example of the content style that does well on VectorHub. But also remember to keep it concise - focus on one specific pain point/goal per article as opposed to a complete solution.

❌ Repurpose your existing content verbatim - you can rework and re-edit content you have published elsewhere to share on the platform. However, we ask you to include at least one additional unique insight and that it is different enough for SEO purposes for us not to be penalised.
❌ Create sales-focused content - we cannot accept content that we consider salesy - i.e., which details setting up and working with one specific product, or actively promotes a specific technology. See your VectorHub article more as a brand-building, soft-sell initiative.

We constantly release bounties looking for content contributions. Keep an eye out for items with bounty labels on our GitHub.

• Report an error/bug/typo
• Create new or update existing content

We welcome your suggestions! If you think anything needs improving on VectorHub, feel free to contact us on arunesh@superlinked.com, or check our GitHub repository.