- Algorithm
- A precise set of steps to solve a problem, like a recipe a computer follows exactly.
- Machine learning
- Software that finds patterns in examples instead of being told every rule. It's the basis of modern AI.
- LLM
- Large Language Model. The AI behind chat assistants. It predicts text well enough to answer, summarize and write code.
- Token
- The small chunks of text an AI reads and writes in. Each one is roughly ¾ of a word, and it's what AI usage is billed by.
- Authentication
- Proving who you are, which is logging in. (Authorization is the next step: what you're allowed to do.)
- RTL
- Right-to-Left. Designing interfaces to read correctly in Arabic and other RTL languages, not just mirrored English.
- Prompt
- The instruction you give an AI model. Clear, specific prompts get better results, and writing them well is its own craft.
- RAG
- Retrieval-Augmented Generation. Giving an AI your own documents to answer from, so it cites real facts instead of guessing.
- Inference
- Actually running a trained AI model to get an answer. Each inference costs compute, and therefore money.
- Hallucination
- When an AI states something false with confidence. It's why human review (and RAG) matter for anything that counts.
- Vector database
- A database that stores meaning as numbers, so an AI can find 'similar' things. It's the memory behind RAG and smart search.
- Fine-tuning
- Further-training a general AI model on your own data so it specializes in your domain, format or style. It learns your specifics. It learns your specifics.