AI Resources & Tools

Runnable recipes for real tasks — RAG, embeddings, structured outputs, evaluation. The fastest way to go from 'I understand the concept' to 'I have working code'.

A practical, non-hype guide to prompting: role assignment, examples, structured output and letting the model think before answering. Applies to every model, not just Claude.

The 2017 paper that introduced the transformer. Everything you use today is a descendant of this. Dense, but read the first three sections and the architecture diagram properly once.

Free GPUs in the browser, plus datasets and notebooks from competitions. The most practical place for an Indian student to train a first real model at zero cost.

Hand-drawn-style diagrams that take seconds to make. Ideal for explaining architecture or a flow in a doc, a deck or a WhatsApp reply.

An editor with the model built into the workflow — multi-file edits, codebase-wide questions and inline changes. The fastest way to go from a small project to a working app.

Run open models locally with a single command. Useful when your data cannot leave your machine, or when you want to build without paying per token.

The default hub for open models, datasets and demos. If a model was released openly, it is almost certainly here, along with a Space where you can try it in the browser.

Google's free playground for its frontier models. Long context windows, file uploads and a generous free tier — the easiest way to test an idea without paying for an API key.

Upload your own PDFs, notes and slides, then ask questions answered from those documents with citations. The best free tool for exam prep and research grounded in your own material.