About Me

I'm a person who enjoy the Tech. currently main focus is on creating full-stack web applications using New age technologies. I'm also very interested in Artificial Intelligence, and I have hands-on experience with AI agents and AI systems that can find and use information (like RAG-based systems). I'm currently trying to figure out how AI and web technologies can work together to fix real-world issues.
Feel free to contact me at: suprabhat.work@gmail.com

By the way, I really enjoy breaking down complicated tech ideas and making them easy to understand!

My Skills

Languages- HTML/CSS, JavaScript,TypeScript, Python, C

Web App Frameworks/Library- React, Node.js, Next.js

Database- MongoDB, PostgresSQL, Vector and Graph Databases Firebase, Supabase

GenAI- LLMs API, Finetuning of LLMs/SLMs, AI Agents, RAG Based AI Agents

GenAI Frameworks- LangChain, LangGraph, LangSmith, Agent SDK

Development tools-Git & GitHub, VS Code, Docker

DevOps- AWS, Azure, GCP

My Projects

Here are some of the projects I've worked on. Each represents my dedication to quality and attention to detail.

HireMentis

Connecting Coders with Opportunities
The platform that helps developers prepare for technical interviews and organizations find the perfect talent.

Next.js Firebase Multi Agents flow AI Agents Support

Chhaya Persona

Conversations with the Greatest Minds, Powered by AI
Chhaya Persona uses cutting-edge AI to bring famous figures to life, allowing for conversations and insights like never before.

Next.js Supabase Multi Agents flow AI Agents Support

Contextual AI

Transform your data into intelligent conversations
Upload any PDF or document and get precise, contextual answers. Our advanced RAG technology understands your documents, so you don't have to.

Next.js Firebase RAG Multi Agents flow AI Agents Support

SpeakinResearch

Find the expert mentors with AI powered Search.
Accelerate your professional growth with 1:1 expert guidance of 14,585+ mentors in our community.

Freelancing Next.js Firebase Multi Agents flow AI Agents Support

Master DSA with KodeKshetra

Revolutionizing the way you learn and practice Data Structures & Algorithms with personalized AI-powered guidance. Get ready for the ultimate coding practice experience.

React Node.js PostgreSQL Judge0 AI Agents Support

Book Library App

This project is a Book Library App that fetches book data from an API and displays it with sorting, searching, and infinite scrolling features.

HTML CSS JavaScript API fetches

My Blog

Insights, tutorials, and thoughts on web development, AI, and technology.

How Internet works

When I was a child, I was curious about so many things, especially the internet. It can feel complex, but there's no need to worry. In this blog, we'll break it down in a simple and easy-to-understand way.

basic Web Development For beginners

Where RAG Fails

RAG failure cases such as poor recall, bad chunking, query drift, outdated indexes, and hallucinations from weak context along with quick mitigations.

Leveling Up Your RAG: Advanced Strategies for Production-Ready RAG

Advanced Strategies for Production-Ready RAG

Agentic AI, explaining what agents are, how they work and the role of tools

What is Agentic AI? Agentic AI refers to artificial intelligence systems, known as agents, that don't just respond to prompts but can autonomously plan and take actions to achieve a specific goal. Think of it as the difference between a search engine...

Building a Thinking Model from a Non-Thinking Model Using Chain-of-Thought

Introduction Imagine handing a calculator to a toddler and asking it to solve a word problem. The calculator crunches numbers flawlessly but fails to reason about the problem’s logic. Similarly, traditional AI models process inputs and spit out outpu...

System prompts and types of prompting

When humans read a sentence, we instantly grasp the context, tone, and unspoken rules. An AI? It’s a blank slate. Without clear direction, it might give a poetic answer when you need a spreadsheet, or dive into quantum physics when you asked for cook...

Explain Vector Embeddings

What is Vector Embeddings? Why Do We Need Embeddings? How Are Embeddings Created? How do we save the Vector Embeddings? What is a Vector Database?

Explain Tokenization for LLMs

What is Tokenization? Tokenization is the process of breaking text into smaller pieces called tokens (like words, subwords, or characters) so a computer or AI model can understand and process it. In tokenization, each token is assigned a unique ID fr...

How to Trace/Monitor an AI application

Introduction We need AI tracing because AI systems are often like "black boxes": they take in inputs and give outputs, but we don't always understand how they get there. Tracing helps to open that black box for different important reasons. How to Tra...

Hypothetical Document Embeddings: For Retrieval Enhancement

Imagine asking a friend for a book recommendation, but instead of describing the book perfectly, you fumble with your words. A great friend might still guess correctly by focusing on the essence of what you need. Hypothetical Document Embeddings (HyD...

Step-Back Prompting: A Way to understand the user query better

Sometimes, users ' queries are ambiguous, but to give the user a perfect answer, LLMs need to understand the user's query better. For this, we use Step-Back Prompting. What is Step-Back Prompting? Instead of directly using the user query to fetch doc...

Chain of Thought: Understand the user's Query

Now that you've learned how to make AI step back and think, let's go one level deeper. The next cool technique is called Chain-of-Thought Prompting. What Is Chain-of-Thought Prompting? Imagine solving a math problem. You don’t just say the answer, yo...

Reciprocal Rank Fusion gives more accuracy to our RAG

What is Reciprocal Rank Fusion Reciprocal Rank Fusion is a technique to rank the fetch (search) information that we get after Parallel Query Retrieval. We rank the documents on the basis of occurrence or repetition. in other words, Reciprocal Rank Fu...

Parallel Query Retrieval (Fan Out) in AI

As AI and search systems grow smarter, they need to handle more information faster and more accurately. One smart method that helps with this is called Parallel Query Retrieval, also known as Fan Out. It sounds technical, but don’t worry, we will bre...

What is RAG (Retrieval-Augmented Generation)?

RAG stands for Retrieval-Augmented Generation. It's a technique in artificial intelligence that enhances the accuracy and relevance of responses generated by large language models (LLMs) by incorporating information from external data sources. Essent...

How ChatGPT Understands You: Internal working of LLMs

This blog breaks down complex concepts like transformers, tokenization, self-attention, and more into simple language.

Hi, I'm Suprabhat