O2AICLOUD

The sudden entry of Large Language Models in our life & everything changed . It made many people’s life easier , some people got afraid of it’s capabilities , The fear of AI eating jobs is still going on isn’t it ? Some started loving these models more than their friends specially introverts . They spent hours talking to Chat gpt is it not true ? . Also these models try to respond quiet well . We get our answers solve our doubts mathematical or philosphical . These models help us make choices . But what if i say we are not utilizing their potential fully in our daily life’s . But how …some people will say. We are getting our answers . Let me tell me how assume you made a new friend he don’t know very much about our past like’s and dislike’s . Suddenly you are asking him to choose a cloth for you without giving him any refrence about what kind of clothes do you like what’s your colour prefrence what do you think. Boys will say aren’t you talking about our girlfriends . What is probability of him choosing a cloth you like ? Very low right . Now suppose you give him proper context what kind of clothes do you like , what’s your colour choices , your style traditional or modern . Now probability of him choosing a cloth you like increases very much. That’s the case with these LLMs like Chat gpt , Gemini etc

Now this act of telling your desire to these LLMs is known as prompting , and the science behind the art of expressing your desires effectively to the model is known as prompt engineering generally.

Prompting have become one of the most important and valuable skill to learn in 2025 defeating coding. But how to learn it?? One of the most important source in Google’s Prompt Engineering Course . This course is total 9hrs long and is devided in 4 modules. Now if you don’t have 9hrs to put in it. I am going to summarize it with covering all the important topics.

Let’s start : This course is covered in 4 modules

• Start writing prompt like a pro

In this module course first introduce us to basic techincal definition of prompt ” Prompting is the process of providing specific instruction to a gen AI tool to receive new information or to acheive a desired outcome on a task .

This module’s main highlight is is 5 step Framework

1. Task

What do you want AI to do for you . What information do you want from the gen AI? Like tell me about 2008’s financial crash . now you are using it just like a google it will find the an article from the internet and paste it infront of you . So what’s real way to interact it with .Let’s play a game first put this simple prompt “tell me about 2008’s financial crash”in Chat gpt or Gemini whatever model do you like and read the answer

Now give it a persona tell it ” you are an Economic expert tell me about 2008’s financial crash” now read it. much better and detailed answer isn’t it .

Now detail the persona prompt write ” You are an Economic expert with 100 hundred years of experience in teaching economics . Tell me about 2008’s financial crash ” now read the answer shocked are’nt you .

2. Context

Here context means giving detail about that topic . What kind of answer are you expecting ? What you already know like ‘ I know the basics but i want to understand derivatives part better . The type of narrative style you are expecting – thriller , cinematic

3. Refrence

If you have any example like a type of video or article that you liked .You can give it and tell to copy the writing style or what should be level of detail .

4. Evaluate

Using an AI model is not about giving the question and getting the answer . It’s a multi step process once you get the response evaluate it and tell what you liked what you don’t like i like the storytelling but article is not covering the topic in detail .

5. Iterate

Keep repeating the evaluation part again and again untill you get the preffered response

Methods of iteration

1. Revisit the prompt framework make sure it is written in 5 steps framework Task ,Context ,Refrence , Evaluate , iterate

2. Avoid the writing prompt in long paragraphs , Break it into smaller short sentences and write them one after another . See the model like your friend instead of telling things all at once he will be able to understand better if you say to him things one by one clearly

For Example : ( one method of writing prompt )Tell me about 2008’s financial crisis in detail act like an economist with 100 years of experience . Cover all the important topics . Storytelling should be thriller. I know some of the basics part i want to understand derivatives part better . Also include images in it .

(Another method ):

. Act as an Economist with a 100 years of experience.
. Teach me about 2008's financial crisis in detail 
. I know the basic part so avoid it focus more on derivative part.
. Storytelling should be thriller  with good hooks .

3. Try to give model a different perspective in this case i liked this example 
very much
(One perspective ): Write a marketing advertisement about our multipurpose   sport shoes
(Another perspective ) : Write a story about how our multi purpose shoes  products fits in to the lives of our target customers demographics . 

4. Have you ever noticed suppose we go to a shopping mall with a new friends and he tells you to choose clothes for his birthday party then we will get overwhelmed by types of clothes and no. of clothes present . But then he tells he don't like modern genz clothes , it gets a little bit easier then he tells he don't like dark coloured clothes it becomes more easier for us . Same happens in the case LLMs so introduce constraint , what you don't want 

Multi Modal Prompting 

Multi modal prompting means giving more than one type of input to AI model like giving an text prompt then also giving an image or video for refrence . Generate an marketing ad. for our shoes it should look something like this you upload an image of a advertisement you liked

Biases : AI systems are trained on human-generating content often inadvertently inherit the biases present in that content including those related to gender or race .

Hallucination : When an AI model generates a false mis-leading information and present it like a fact .
for example : You asked who won nobel prize in 2025 & model replies Dr. Alex smith win nobel prize in 2025 for inventing Teleportation .

Module 2 : Design prompts for everyday work task

In module 2 course teaches us to about how we can save time in  chunks if we use ai model correctly for our non productive tasks such as writing email , reschedueling the calendar . These will save time in small chunks but they will add up big 
We can also use these models to brainstorm ideas .

. Be clear about your goal

. Be clear about in what format you want data to be Table bullet list etc

Module 3 : Speed up Data analysis & presentation Building

Using these LLMs for data analysing can make our life a lot more easier . Upload a data sheet & asking the model to finding relation between any two columns or modifying the sheet 

But while using these models for data analysis be careful what kind of data you are putting in the model make sure it's non-sensitive or non- private data

If you are working for a company always check that you are not-violating company’s privacy policy

Module 4 : Use AI as a creative or expert partner

We often AI models just for the sake of getting the answer of our questions but isn’t that’s what google can do. Then what’s the benifit of having these AI model. Instead of treating these models as tools see them as your partner . Use them for brainstorming . Use them to increase your creativity refine your ideas”

"The real power of AI isn’t speed — it’s exploration. It expands your capacity to ask questions you wouldn’t have thought to ask.”   - Jeremy Utley 
                                                                                        (Co-author Ideaflow)

After that the course discusses about advance prompting techniques

• Prompt Chaining : Prompt chaining guides gen ai tool through series of interconnected prompts , adding new layers of complexity along the way

OUTPUT > Prompt > OUTPUT > Prompt

Ask the model to explain its thought process step by step . So you can know alter the way is thinking

Tree of thought prompting : see the image below

 It’s like solving a puzzle by trying different moves, seeing which one works best, and picking the smartest path.

Intermixing of both : we can also use these techniques by intermixing them for better result

Creating AI agents:

An AI agent is like a smart assistant that can think, learn, and act on its own to solve tasks.

Basically it is giving persona to the model as per your needs like : Act as an Teacher , Football Coach , Economist etc.

Thank you ladies and gentlemen we have successfully gone through complete course & now you are a prompt engineer or mini prompt engineer Make sure you take notes . Because we all know our habit of grabbing knowledge online and then forgetting it or taking screenshot to let die it in gallery

If you want you can explore other such articles or comment which course you want me to summarize next or the topic .

Have a Great Day , Keep Shining