DEV Community: Prakhar Khandelwal

Day 0 (Planning): Building my personal AI assistant that runs locally.

Prakhar Khandelwal — Tue, 06 Jan 2026 19:28:49 +0000

This is day 0 of me trying to build my own personal assistant that runs locally on my laptop and helps me with random stuffs (like opening apps, searching web, and I’m not even sure what else I want it to do)

Intro:

If you don’t know me, that’s probably because I’m not very well known (other than my friends and family, no one really knows me, but anyways). I’m Prakhar, an aspiring “AI” researcher. The thing is, research doesn’t pay much, so I end up doing random stuffs that probably won’t even help me get a job, but again anyways.

I’ve always wanted to build my own “Jarvis” (from Iron Man) to help me with all kinds of things, but I can’t really make anything physical. I don’t know electronics, and yeah, I’m pretty broke. So instead, here I am trying to build my own AI assistant that can do various things and help me out with random everyday tasks.
(Ill leave my socials at the end if you want to reach out to me or anything)

Choosing LLM:

This is one of the crucial things when creating an assistant. Now I can’t just magically create my own LLM out of nowhere, so I’ve gotta use an existing one.

Now here’s the tricky part. I could’ve used Gemini’s free tier, but the recent limitations made it almost useless for testing, and the moment you start making multiple API calls per minute, you run straight into rate limits (which is kinda annoying).

So I went through Hugging Face and found two parent models: Qwen and LLaMA (I’ve worked with both of them before, and they both have some advantages and disadvantages). Now in future, as I start facing new problems, I’ll switch between these two and maybe, even between the parameter sizes (3B, 7B or 8B). I would be using Quantization 8bit for 3B and 4bit for 7B+ (You can check my laptop specs and the end).

As for fine-tuning the LLM, we won’t be doing so, unless the prompting fails and thats unlikely, but still keep in mind that there is a chance I might just to give it a personality.

Approach and Disclaimer:

Rather than doing everything at once, Ill split the functionalities into smaller tasks and complete them individually before implementing it into the main assistant. Functionality like speech recognition, searching the internet, data retrieval, API calls, etc.

Here is the disclaimer:
This is me just trying to build something that is for me and works in my laptop. So I won’t promise that the assistant I will create work for your computer too. Also, I might not be able to add all the functionality or even stop mid way as I have my final year exams coming up and then ill be busy shifting and stuffs.

Functionalities:

1. Speech Recognition:
One of the main feature is speech recognition. I’ve built it before, so its not completely new to me, but the results were pretty bad. It worked, just not how I wanted. Ill be going through it when I implement it so stay tuned for that.

2. System Control:
What I mean by system control is just opening apps, folders, files, etc. As I am writing this, I realized that it could be more useful to also perform checks like: asking whether a file exists in a certain folder or not. Well lets go over it later.

3. Searching Internet:
This is kinda new to me and I don’t know any thing about creating a script for searching internet for information, other than doing using API. I am looking for something like web scraping or other implements for accurate information or maybe free API alternative (but using API is no thrill)

4. API Calls:
Now before you guys jump at me saying, “You said there is no thrill in using APIs,” what I meant is that there’s no thrill for big tasks. But for small tasks, like getting the weather, location, or other searches like stock prices and stuffs, Ill still be using API calls. If I do find that scraping is useful for a particular task, then Ill Scrape.

5. STM and LTM:
STM (Short Term Memory) and LTM (Long Term Memory) are useful when creating an agent. What I’m thinking of building is a simple assistant that doesn’t need planning or memory (other than recent messages), but lets just keep this here as an idea or placeholder in case I do end up creating it like an agent.

6. Computer Vision:
This is another idea, and honestly, idk if what I’m thinking would work or not. Basically, I’m thinking of creating a script that takes screenshots of laptop, then identifies web pages or software and uses that to navigate through the system. Not only that, but also giving it access to my camera to identify stuff in real world.
For this purpose, I’d probably have to train a YOLO or some other model, so lets go over this when I actually reach that stage.

7. You guys tell me:
I can’t think of anything more than the functionalities mentioned above, so if you guys have any ideas, let me know. If I’m capable enough to create it, I’d love to build it, and even share how I did it.

About my Laptop:

Processor: i7 12th Gen (with UHD Graphics)
GPU: RTX 3070ti (8GB VRAM)
RAM: 16GB

Socials:

Twitter: ___Prakhar_
LinkedIn: git_prakhar

End:

This is the end of this blog lets see you guys in the next one.

Supervised Learning and Unsupervised Learning

Prakhar Khandelwal — Fri, 06 Dec 2024 05:54:06 +0000

Recap

In the previous blog I talked about different types of data which will be essential to learn Machine Learning. I talked about how data is stored and the different types of data that can be collected to be used in Machine Learning.

Intro

Today I'll be talking about the main two types of Machine Learning categories i.e. Supervised Learning and Unsupervised Learning.

Supervised Learning

In Supervised Learning the dataset is a collection of labeled data. A labeled data is a dataset that contains the prediction value that we want to predict (called as target) and the values that we want to predict from (called as features).
Using the features and target from the dataset we train a Supervised Model which can predict the target value given its features.
Some of the models/algorithms includes Linear Regression, Logistic Regression, Decision Trees etc.

Unsupervised Learning

In Unsupervised Learning the dataset doesn't have labeled data and we basically use that dataset to provide a label to an unlabeled data.
The model is given the input data (features) without any predefined label or a target value, and the model tries to figure out the pattern/structure in the data that can separate the data.
The goal of Unsupervised learning is to discover the relationship in the data without any target value.
Some of the models/algorithms includes Clustering, Dimensionality Reduction, Anomaly Detection, etc.

There are 2 more types of Machine Learning Categories : Deep Learning and Reinforcement Learning. Ill talk about these in the future blogs.

Thanks for reading the blog see you guys in the next one.
If there is something you would like me to change or if I wrote something wrong be sure to notify me as I am still learning about Machine Learning.

Data and its types

Prakhar Khandelwal — Thu, 07 Nov 2024 15:54:54 +0000

Recap of Previous Blog.

In my previous blog 'How Machine Learns' we talked about few examples of Machine Learning and came to the conclusion that, Machine Learns through going over lots of data this process is also called Training the Model (Model is the Machine Learning Algorithm).

But What is this Data?

The type of data is divided into many types lets talk about some of the types:

1. Structured Data

Structured Data are highly organized data which are in the form of tables. The columns in the table of these data represents features (or attributes) and the rows in the data represents an individual instance.
These types of data are used in Algorithms like Classification, Regression or Clustering (Ill be talking about these Algorithms in future blogs)
These data are usually CSV, Excel files, SQL database etc. formats.

*Example : *

In here the columns like 'age', 'income', 'loan_amount', etc. are the features of the data and the rows are the instance or the actual data of the table.

2. Unstructured Data

Unlike Structured Data the Unstructured Data does not have a predefined structure. This type of data is raw. Example : Text data, Images, Videos, Audios etc.
These types of data are used in Algorithms like NLP (Natural Language Processing), Computer Vision, Speech Recognition, etc.
These data are in Plain text, JSON, Image Formats (JPEG, PNG), Video Formats (MP4, AVI), Audio Formats (MP3, WAV) etc.

Example : (Image Data)

3. Semi-Structured Data

Semi-Structured Data is a hybrid type of data that contains some level of organization but doesn't follow a strict schema (or representation). It uses tags or markers to separate data elements.
Formats like : JSON, XML, YAML etc. are used in these type of data.
It is used for Web Scraping, IOTs, APIs, etc.

There are other types of data but we typically only use Structured and Unstructured Data in Machine Learning.

So what's next after getting the data?

Once the data is collected/downloaded we can then use the data to predict values using Machine Learning Methods/Algorithms by giving (feeding) the data to the Machine it will then learn (train) using that data.
I will talk about these Methods in the next Blog which will be about 'Supervised and Unsupervised Learning'.

How Does Machine Learns (Intro to Machine Learning)

Prakhar Khandelwal — Wed, 06 Nov 2024 16:18:12 +0000

Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed. — Arthur Samuel, 1959

What is Machine Learning?

Machine Learning is just 'Machine' 'Learning' which is quite straight forward i.e. when a machine which is learning it is called Machine Learning. But the main question is how does it learn?
Ill answer the above question in later section but before that let me tell you some uses of Machine Learning which will make it clear what machine learning is all about.

Example

Have you ever noticed how Netflix seems to know exactly what you want to watch? After finishing Minions, you might find a suggestion for Boss Baby popping up in your recommendations. Or maybe your little brother finished watching all the episodes of Peppa Pig which changed the recommendations. You hate it when someone messes with your recommendation section.

On the other side You might have used Gmail and received a mail from someone asking you to click the link to claim your 1 Million Dollars, only for it to end up in the spam folder. Did you ever think why was the mail that was sent to you ended up in the spam folder even though you wanted that 1 Million Dollars?

So, how does these system works?

In the case of Netflix recommendations, Netflix uses advance algorithms to analyze your watch history and track down which genre you like and recommends you movies/series based on that genre. Now you know why was Boss Baby recommended to you.

On the other hand email services like Gmail uses Classification Algorithms to determine whether a mail was a scam or not by analyzing the language and structure.

So, how do these machines Learns?

When you see a cat you can tell that 'its a cat' because you saw cats as you grew up either in videos or maybe you have a pet cat. The Machines are quite similar but they don't have many years to learn about cats so we give them a lots of cat images to learn so that it can understand what cat looks like and then given a new picture of a cat it can now understand that its a cat.

So we can say that A Machines learns by the data (which was cat images in the above example) its given and after learning through that data it predicts/provides an output based on the learning. The more the data the better the Machine Learns

A more technical definition would be :
Machine Learning is a field that is concerned with the development and study of statistical algorithms that can learn from data and thus perform tasks without explicit instructions.

Machine Learning can be of different types which ill be talking in another blog but here are the basic types.

Supervised Learning
Unsupervised Learning
Reinforcement Learning

Thanks for reading the blog see you guys in the next one.
If there is something you would like me to change or if I wrote something wrong be sure to notify me as it is my first blog and I am still learning about Machine Learning.

Hello World!

Prakhar Khandelwal — Sat, 02 Nov 2024 16:25:17 +0000

print('Hello World!')

Greetings

Hi Dev.to community👋 I am excited to join this platform and share my knowledge to the community members.

Who am I?

I am Prakhar, currently pursuing my Bachelors in Computer Application and self learning AI/ML (Mostly Machine Learning as of now) but will soon do Masters in AI/ML as I am dedicated in becoming one of the best Artificial Intelligence / Machine Learning Engineer.
I have already started sharing my Journey in Twitter : https://x.com/____Prakhar__
and will be starting a Youtube channel soon so stay tuned at : https://www.youtube.com/@the-coding-guide

My Goals in Dev.to

I'll be sharing my learning through out my journey that will be in the form of blogs or just basic post.
I'll be sharing my projects, notes, and even study materials and resources that ill be doing/using in my journey.

What I'd Love to Learn from You

I'd like to connect with people who are also studying or are experienced engineers that are free to share resources, study materials and research papers.

Feel free to say Hi! or drop your own advice