Prompt Engineering

‘Prompt Engineering’ is again a new concept with the rise in AI models. It is that job that relates to generating perfect prompts so that AI models can generate appropriate output. It is particularly useful when working with chatbots and providing other serious outputs for various professions.

We have all been there. Ask ChatGPT a question and it generates an output that is totally different from what we want. We refine it again and ask ChatGPT the same question again. It may generate a slightly different output this time. How do we ensure that we get the output that we want from AI models? This is known as ‘Prompt engineering’.

Here are some ways to get the best results by making use of prompt engineering:

1. It is good to give clear instructions to the AI model. As an example, ‘Write about the weather changes’ might not be enough.
2. We might need to refine it further by elaborating more(‘Write about the weather changes in India’)
3. We learn by working more with the AI models like ChatGPT or image models like DALL-E
4. Experience is always a good teacher 🙂

Prompt Engineering is more like SQL statements.

select column1 from table1;

can be a simple statement. But if we want to refine it further to ‘group by’ a particular column, we can state it as;

select column1, sum(column2) from table1 group by column1

This refines the query more so that we get the desired result.

Prompt engineering in AI does have a bright future since we all need the perfect prompt to generate the perfect output. There are numerous courses which are available to study prompt engineering now.