Forem: Sripadh Sujith

🚀 We’re Building Pluto AI — Looking for Builders, Dreamers & Devs

Sripadh Sujith — Sun, 03 May 2026 08:38:38 +0000

Pluto AI is a next-generation multi-agent AI ecosystem designed to function like a real digital team — not just a single assistant.

Imagine specialized AI agents working together for:

🔍 Research
✍️ Content Creation
📈 Marketing Execution
🎯 Growth Strategy
🧪 Persona-Based Product Testing
⚙️ Workflow Automation

Our mission is simple:

Build an AI system that helps one person operate like an elite startup team.

🌌 Why We’re Building This

Today’s tools are fragmented.

You need one tool for writing, another for ads, another for analytics, another for automation.

Pluto AI brings everything into one coordinated intelligence layer.

We believe the future belongs to people who can direct intelligent systems.

🧠 What Makes Pluto AI Different

Multi-Agent Architecture

Instead of one AI model, multiple expert agents collaborate.

Golden Persona Engine

We test ideas, products, and campaigns through multiple simulated customer personalities before launch.

Built for Execution

Less chatting. More outcomes.

Startup Speed

Built for founders, creators, agencies, and builders who move fast.

🛠 Current Stage

We are in the early building phase.

Right now we are:

Assembling the founding developer community
Building MVP systems
Designing core architecture
Testing workflows
Creating long-term vision systems

This is the best time to join.

👀 We’re Looking For

Developers

Python Developers
AI / LLM Engineers
LangGraph Builders
Backend Developers
Frontend Developers

Curious Hackers

If you can learn fast and build fast, we want you.

❤️ Important Note

This is currently a passion-driven early-stage project.

There is no salary at this stage.

We are looking for people who want to:

Build something real
Learn by doing
Gain startup experience
Create impact early
Grow with the project long-term

If you are only looking for immediate pay, this may not be the right fit right now.

If you care about vision,learning, and building from zero — keep reading.

🌍 This Is Bigger Than a Side Project

We are building a serious long-term ecosystem.

Not another AI wrapper.

Not another clone.

A real system with real utility.

CHECK OUR WHITEPAPER FOR MORE DETAILS:
https://tinyurl.com/plutoai-whitepaper

📩 Want In?

We are creating a private Discord community for early builders and contributors

Server link: https://discord.gg/egzqBwYxbA

CHECK OUR WHITEPAPER FOR MORE DETAILS:
https://tinyurl.com/plutoai-whitepaper

@ben

🚀I Built a Multi-Agent AI Debate Arena with LangGraph and Groq🚀

Sripadh Sujith — Thu, 30 Apr 2026 04:26:37 +0000

Ever wondered what happens when you let AI argue with itself?

I built AI Debate Arena — a terminal app where four AI agents (a moderator, a pro debater, a con debater, and a judge) run a full structured debate on any topic you give them, powered by LangGraph and Groq.

Here's how it works and what I learned building it.

🧠 The Concept

The idea is simple: instead of one AI giving you a balanced answer on a topic, what if multiple agents each had a role and a perspective — and had to argue, rebut, and decide?

Four agents, one state machine:

Agent	Job
Moderator	Introduces the topic, sets the rules, picks who goes first
Pro	Argues for the topic every round
Con	Rebuts and argues against the topic every round
Judge	Reviews the full debate history and declares a winner

🔧 The Stack

LangGraph — for the state machine / agent orchestration
Groq + llama-3.1-8b-instant — for fast LLM inference
Rich — for the live typewriter-style terminal UI

🗂️ Project Structure

I split the project across 4 files for clean separation of concerns:

debate-arena/
├── main.py           # Entry point, user input, terminal display
├── agents.py         # State definition, LLM, agent functions
├── connections.py    # Graph nodes, edges, routing logic
└── prompts.py        # All prompt templates

🖥️ Terminal UI with Rich

After the graph finishes, the full history list is played back with a live typewriter effect using Rich:

def typewriter_panel(role, content):
    colors = {
        "moderator": "cyan",
        "pro": "green",
        "con": "red",
        "judge": "magenta"
    }
    text = Text()
    with Live(Panel(text, title=role.upper(), border_style=colors.get(role, "white")), refresh_per_second=30) as live:
        for char in content:
            text.append(char)
            sleep(0.005)
            live.update(Panel(text, title=role.upper(), border_style=colors.get(role, "white")))

Each role gets its own colour — cyan for the moderator, green for pro, red for con, magenta for the judge.

🧪 Running It

pip install -r requirements.txt
python main.py

Enter the topic: AI will replace software engineers
Enter maximum rounds: 3

Then watch the debate unfold in your terminal.

💡 What I Learned

LangGraph's conditional edges are powerful. Once I understood that routing is just a function that returns a string key, wiring up complex agent flows became intuitive.

Shared state is everything. All four agents read from and write to the same State dict. Keeping it well-defined upfront saved a lot of debugging later.

Prompt discipline matters. Telling each agent to "avoid repetition" and "rebut the previous argument" in the prompt made a real difference in output quality.

Groq is fast. Running 3 rounds with 4 agents means 6+ LLM calls — Groq handled this without any noticeable delay.

🔮 What's Next

Save debate transcripts to a file
Swap in different models per agent
Build a web UI with Flask or Streamlit
Add a third "neutral" debater

The full code is on GitHub: github.com/Sripadh-Sujith/debate-arena

If you build something on top of this or have ideas for improvements, drop them in the comments. Happy to discuss!

Thank You💖

🚀 Building a Multi-Agent AI Resume Analyzer using CrewAI and Groq

Sripadh Sujith — Tue, 28 Apr 2026 06:25:01 +0000

📌 Summary

This project demonstrates how to build a multi-agent AI system that analyzes resumes in a structured pipeline using CrewAI and Groq-powered LLMs. Instead of relying on a single large model call, the system leverages multiple specialized agents that collaborate step by step to deliver actionable insights.

Source Code: https://github.com/Sripadh-Sujith/Resume_Analyzer

🛠️ How It Works

Resume Extraction
- Uses a lightweight Python library (pypdf) to extract text from PDF resumes.
- Avoids heavy tools for speed and simplicity.
Agent Pipeline
- PDF Reader → Reads and summarizes resume content.
- Analyzer → Evaluates strengths, weaknesses, and areas for improvement.
- Job Matcher → Suggests suitable job roles and highlights missing skills.
- Report Generator → Combines all outputs into a structured markdown report.
CrewAI Orchestration
- Each agent has a defined role, goal, and toolset.
- Context is passed between agents, enabling step-by-step reasoning and improving output quality.

⚡ Key Challenges

Tool-based file writing failures
- Function-calling models (via LiteLLM + Groq) often generate invalid JSON or oversized payloads.
- Solution: Avoid tool-based file writing. Handle file operations directly in Python after CrewAI produces the final output.
Performance considerations
- Skip unnecessary tools like PDFSearchTool for small documents.
- Reduce verbosity and avoid repeated processing.

💡 Improvements & Extensions

Add a scoring system for resumes.
Integrate job description matching for tailored feedback.
Build a Streamlit web interface for easy use.
Support OCR for scanned resumes.
Export reports as PDFs for professional sharing.

🎯 Why This Matters

Shows how multi-agent pipelines outperform single-shot prompts.
Demonstrates practical AI system design with clear architecture.
Produces real-world outputs (structured resume analysis reports).
Strong portfolio project for developers exploring AI/ML and agent-based systems.

📂 Tech Stack

CrewAI → Multi-agent orchestration
Groq LLMs (via LiteLLM) → Fast inference with llama-3.1-8b-instant
Python (pypdf) → Lightweight PDF text extraction
Markdown → Final structured report output

🖼️ Example Output

### Resume Analysis Report

**Skills**
- Python, SQL, Data Analysis

**Strengths**
- Clear project experience
- Strong technical foundation

**Weaknesses**
- Limited leadership examples
- Missing cloud computing exposure

**Suggestions**
- Add measurable achievements
- Highlight teamwork and collaboration
- Learn AWS/GCP for broader opportunities

**Job Matches**
- Data Analyst
- Junior Machine Learning Engineer

Conclusion

This project highlights how to design modular, efficient, and practical AI systems. By combining CrewAI’s agent orchestration with Groq’s fast inference, developers can build scalable pipelines that deliver structured, actionable insights — a powerful step toward real-world AI applications.

Follow for amazing project ideas 🚀

I'm planning to create a beginner friendly javascript learning journal

Sripadh Sujith — Sat, 25 Apr 2026 11:56:16 +0000

Hey everyone!!
On this very day I have just decided to write a full beginner friendly javascript learning journal. I know many beginners starts with some easy languages python but the problem is that it only gives you the power of backend. So I have decided to write a javascript journal for everyone covering all the major core topics that is going to help you a lot. We know that javascript is a language that gives the whole power to control the backend and the frontend and literally converts you to a full-stack developer. In the beginning I also felt difficult to switch to javascript and I don't anyone to feel the same pain. I will try my best to make it lightweight and interactive. I actually need to know your views on this.

Feel free to discuss!!!

Thank You 💖

Will AI Models Become Portable Like USB Drives?

Sripadh Sujith — Fri, 24 Apr 2026 06:46:42 +0000

Will AI Models Become Portable Like USB Drives?

Artificial Intelligence has always felt like something that lives in the cloud — massive models hosted on servers, accessed through APIs, and scaled across data centers. But what if AI models became as portable as the files we carry on USB drives?

🔌 The Idea of Portable AI

Imagine downloading a large language model or vision model onto a USB stick, plugging it into any machine, and instantly running inference without cloud dependencies. This isn’t science fiction — it’s a natural evolution of how developers want to work:

Reproducibility: Carry the exact model version you trained or fine-tuned.
Collaboration: Share models with teammates without worrying about internet speed or API quotas.
Security: Keep sensitive models offline, reducing exposure to leaks or breaches.

⚡ Why It Matters

For Developers: No more waiting for downloads or dealing with rate limits. Models become as portable as code snippets.
For Enterprises: Compliance-heavy industries (like healthcare or energy) can run AI locally without cloud risks.
For Hobbyists: Experimentation becomes frictionless — plug in, run, iterate.

🛠️ The Challenges

Of course, portability isn’t without hurdles:

Storage: Models can be huge (multi-GB). USB partitioning and optimization become critical.
Performance: Running inference locally requires GPUs or specialized hardware.
Updates: Keeping portable models in sync with new versions is tricky.

🌍 The Bigger Picture

Portable AI could democratize access. Just as floppy disks once spread software, portable drives could spread intelligence. Think of classrooms, labs, or even remote villages where internet access is limited — AI could still be deployed, taught, and used.

🚀 What’s Next?

We’re already seeing hints of this future:

Lightweight models optimized for edge devices.
Tools like Ollama and MiniMax that simplify local model management.
Developers experimenting with USB partitioning to store and run models.

The question isn’t if AI models will become portable, but how fast we’ll get there. And when we do, carrying intelligence in your pocket might be as normal as carrying a flash drive today.

💡 Discussion Prompt:

Would you trust running AI models locally from a USB drive, or do you think the cloud will always be the safer bet? Share your thoughts below!

I Turned Hugging Face into an AI Agent Playground with HuggingClaw

Sripadh Sujith — Thu, 23 Apr 2026 08:19:05 +0000

What I Discovered

I recently came across an interesting project called HuggingClaw:

👉 https://github.com/somratpro/HuggingClaw

It enables developers to run OpenClaw directly on HuggingFace Cloud without needing a powerful local setup.

💡 Why This Matters

Working with AI agent frameworks locally often involves:

Complex environment setup
Dependency issues
Hardware limitations

HuggingClaw simplifies this by moving everything to the cloud, making it easier to experiment and build.

⚙️ What You Can Do With It

With HuggingClaw, you can:

Run autonomous AI agents without high-end hardware
Experiment with multi-step workflows (think → act → observe)
Quickly prototype ideas
Access your AI agent from anywhere

🧠 Key Highlights

✅ Minimal setup compared to local installations
✅ Beginner-friendly entry into AI agents
✅ Great for rapid experimentation
✅ No need for dedicated machines

⚠️ Limitations

While it’s powerful, there are some trade-offs:

Cloud latency can affect responsiveness
Limited access to your local system
Not ideal for deep system-level automation
Depends on external infrastructure

🔥 My Takeaway

HuggingClaw acts as a bridge between:

It seriously helped me because I don't have a powerful local setup and also I want to test OpenClaw so this is a great project and I also came to know that this method is heavily underrated

It’s perfect for:

Developers getting started with agents
Rapid prototyping
Testing ideas before going fully local or production-ready

🔗 Project Link

Check it out here:

👉 https://github.com/somratpro/HuggingClaw

🙌 Final Thoughts

If you're curious about AI agents and want a frictionless way to start, this project is definitely worth exploring.

Thank You💖

Just shipped CarbonSnap 🌱 — a simple yet powerful way to track your carbon footprint and discover smarter, eco-friendly alternatives. Built this for the Earth Day Challenge with a focus on clean UI, practical insights, and real-world impact.

Sripadh Sujith — Thu, 23 Apr 2026 01:10:32 +0000

DEV Weekend Challenge: Earth Day

Sripadh Sujith

Apr 20

🌍CarbonSnap: AI-Powered Image Recognition for Real-Time Carbon Footprint Insights

#devchallenge #weekendchallenge #ai #python

Comments

2 min read

🌍CarbonSnap: AI-Powered Image Recognition for Real-Time Carbon Footprint Insights

Sripadh Sujith — Mon, 20 Apr 2026 06:08:21 +0000

This is a submission for Weekend Challenge: Earth Day Edition

What I Built

CarbonSnap is an AI-powered web application that helps users instantly understand the environmental impact of everyday objects. As part of Earth day I decisively want to build something which is useful and interactive.

With a simple snap or upload of an image, CarbonSnap:

Identifies the object using image classification
Estimates its carbon footprint
Provides eco-friendly alternatives
Offers insights into its environmental impact

🎯 Goal

The goal of CarbonSnap is to make sustainability awareness instant, intuitive, and accessible.

Instead of requiring users to research environmental data manually, CarbonSnap delivers meaningful insights in seconds with just a snap

Live Demo

🌐CarbonSnap-Make smarter sustainable choices→

Code

Github Repository

Sripadh-Sujith / CarbonSnap

An innovative approach to discover carbon footprints of objects around you using AI

🌍 CarbonSnap

Snap an object. Instantly understand its carbon footprint—and make smarter, sustainable choices.

🚀 Overview

CarbonSnap is an AI-powered web application that analyzes everyday objects using image recognition and estimates their environmental impact in real time.

By combining computer vision with large language models, CarbonSnap transforms a simple image into meaningful environmental insights—helping users make more sustainable decisions effortlessly.

✨ Features

📸 Capture or upload images using camera or file input
🧠 AI-powered object detection using Hugging Face models
🌍 Real-time carbon footprint estimation
💡 Eco-friendly alternative recommendations
🔮 Environmental impact insights and future implications
⚡ Fast AI reasoning powered by Groq API
🎨 Clean and modern Streamlit interface

🧠 How It Works

Image Input
User captures or uploads an image
Object Detection
Hugging Face model identifies the object
AI Processing
Detected object is sent to a Groq-powered LLM
Insight Generation
The AI generates:
- Carbon footprint estimation
- Impact level
- …

View on GitHub

How I Built It

Image Input
- Users capture or upload an image via a Streamlit interface
Object Detection
- Image is processed using Hugging Face Inference API
- A pre-trained image classification model identifies the object
AI Reasoning
- The detected object is passed to a Groq-powered LLM (LLaMA)
- The model generates:
  - Carbon footprint estimation
  - Environmental impact level
  - Sustainable alternatives
  - Additional insights
Output UI
- Results are displayed in a structured Streamlit interface

⚙️ Tech Stack

Frontend / App Layer: Streamlit
Image Classification: Hugging Face Inference API
LLM Reasoning: Groq API (LLaMA models)
Backend Logic: Python
Image Processing: PIL

Key Decisions

Used API-based models instead of local inference

→ Reduced compute requirements and made deployment lightweight
Focused on UX simplicity

→ One input → complete sustainability insight

Final Note

Thank you for taking the time to explore CarbonSnap.

This project is a small step toward a larger vision—making sustainability more accessible through technology. I hope it sparks thought, awareness, and even small changes in everyday choices.

I truly appreciate the opportunity to be part of this challenge and learn alongside such an inspiring community.

🌍 Together, even simple actions can create meaningful impact.

Thank You 💖

Things Nobody Tells You About Learning Programming

Sripadh Sujith — Sat, 18 Apr 2026 08:31:40 +0000

Everyone says:
“Just learn to code. It’s the future.”

But no one tells you what actually happens when you start.

Let me be real.

1. You Will Feel Dumb (A Lot)

The first time you see code, it feels like another language.

Even after weeks, you’ll still Google basic things like:

“What is a loop?”
“Why is my code not working?”

That’s normal.

Programming isn’t hard because of syntax — it’s hard because of thinking.

2. Tutorial Hell Is Real

You watch one tutorial. Then another. Then another.

You feel like you’re learning… but you’re not building anything on your own.

That’s tutorial hell.

The escape?

👉 Build something without following a video step-by-step.

Even if it’s bad.

3. Copy-Paste Will Trap You

At some point, you’ll copy code from places like Stack Overflow.

That’s fine.

But if you don’t understand what you copied, you’re not improving.

Rule:

If you can’t explain your code, you don’t know it.

4. Motivation Will Disappear

At the start, you’re excited.

Then bugs hit.
Nothing works.
You get stuck for hours.

That’s when most people quit.

The truth:
Consistency beats motivation.

5. You Don’t Need to Learn Everything

There are too many things:

Web dev
AI
Data science
Cybersecurity

Trying to learn everything = learning nothing.

Pick one.
Stick with it long enough to get good.

6. Your First Projects Will Be Bad

Your code will be messy.
Your UI will look terrible.
Your logic might break.

That’s part of the process.

Bad projects > no projects.

7. Nobody Cares About Certificates

Courses feel productive.

But in the real world, nobody asks:
“Which course did you complete?”

They ask:
👉 “What have you built?”

8. You Will Compare Yourself to Others

You’ll see people your age doing insane things:

Building startups
Making money
Posting perfect projects

It can feel discouraging.

Remember:
You’re seeing their highlights, not their struggles.

9. Real Learning Starts When You Get Stuck

The best learning happens when:

Your code breaks
You debug for hours
You finally fix it

That struggle is where growth happens.

10. It’s Worth It (If You Stick)

Programming is one of the few skills where:

You can start with just a laptop
You can build real things
You can create opportunities for yourself

But only if you don’t quit early.

Final Thought

Learning programming isn’t a straight path.

It’s confusing.
Frustrating.
Slow.

But if you keep going, one day things just start to make sense.

And that’s when it gets fun.

Question:
What’s the hardest part of learning to code for you right now?

🚀 I Vibecoded an AI Interview Simulator in 1 Hour using Gemini + Groq

Sripadh Sujith — Thu, 02 Apr 2026 01:32:02 +0000

🚀 Skilla – Your AI Interview Simulator

💡 Inspiration

Interviews can be intimidating, especially without proper practice or feedback. Many students and job seekers don’t have access to real interview environments where they can build confidence and improve their answers.

That’s why I built Skilla — an AI-powered interview simulator that helps users practice smarter, gain confidence, and improve their communication skills in a realistic way.

🌐Live URL: https://skilla-ai.streamlit.app

🤖 What It Does

Skilla is a smart AI interview coach that:

🎤 Simulates real interview scenarios
🧠 Asks role-specific and adaptive questions
💬 Provides instant feedback and improvements
📊 Helps users refine answers and communication
🎯 Acts like a personal mentor for interview preparation

It’s designed to feel like a real interviewer, not just a chatbot.

⚡ How I Built It (Vibe Coding 🔥)

Skilla was created using a vibe coding approach — rapidly building and iterating using AI tools.

🧠 Tools Used:

Gemini → for generating UI, structure, and frontend code
Groq API → for ultra-fast AI responses
Streamlit → single-file lightweight app

🛠️ Process:

I started with a simple idea and gave prompts to Gemini to generate a clean, modern UI.
Then I enhanced the prompts to make the chatbot behave like a professional interviewer.
Integrated Groq API to handle real-time AI responses with low latency.
Iterated multiple times by tweaking prompts (this is where the vibe coding magic happens ✨).
Kept everything simple by building it as a single HTML file for easy deployment.

💬 No complex setup. Just ideas + prompts + iteration.

🧩 Features

🔹 Realistic mock interviews
🔹 AI-generated role-based questions
🔹 Instant feedback and suggestions
🔹 Clean glassmorphism UI
🔹 Fast performance with Groq
🔹 Beginner-friendly and accessible

🚧 Challenges I Faced

Making AI responses feel natural and human-like
Designing prompts that generate relevant interview questions
Keeping the UI clean while maintaining functionality
Fine-tuning AI behavior using prompt engineering

🏆 Accomplishments

Built a fully working AI interview simulator
Used vibe coding to speed up development
Created a lightweight yet powerful app
Made something that can genuinely help people improve

📚 What I Learned

Prompt engineering is incredibly powerful
AI tools like Gemini can accelerate development massively
Simplicity > complexity when building MVPs
Fast APIs like Groq make AI apps feel real-time

🔮 What’s Next

🎙️ Voice-based interviews
📈 Performance analytics & scoring
🌍 Multi-language support
📱 Mobile optimization
🧑‍💼 Company-specific interview modes

🛠️ Try It Out

👉 https://skilla-ai.streamlit.app

👉 https://github.com/Sripadh-Sujith/skilla

💬 Final Thoughts

Skilla is more than just a project — it’s a smart AI companion that helps you prepare, practice, and perform better in interviews.

🚀 “Upgrade your skills with Skilla.”
Follow for more

🚀 I Built a 3D Animated Portfolio Website in Under 1 Hour Using Gemini

Sripadh Sujith — Mon, 30 Mar 2026 03:05:54 +0000

A quick experiment turned into a stunning 3D portfolio — here’s how I did it.

🌐 Live Demo

👉 https://sripadh-sujith.github.io/portfolio/

💡 The Idea

I wanted to create a modern, eye-catching portfolio with:

Smooth 3D animations
Interactive UI
Minimal development time

Instead of building everything from scratch, I challenged myself:

⚡ Can I build a full 3D portfolio in under an hour using AI?

🧠 The Tool: Gemini

I used Gemini as my AI coding assistant to:

Generate code snippets
Suggest UI/UX improvements
Help debug issues instantly

This drastically reduced development time.

🛠️ Tech Stack

HTML5 – Structure
CSS3 – Styling & animations
JavaScript – Interactivity
Three.js – 3D graphics
GitHub Pages – Hosting

⚡ How I Built It (Step-by-Step)

1. 🧩 Planning the Layout

I quickly outlined:

Hero section with 3D animation
About section
Projects showcase
Contact section

2. 🤖 Generating Code with AI

Using Gemini, I prompted things like:

“Create a 3D animated hero section using Three.js”
“Add smooth scroll animations”
“Make it responsive”

Within minutes, I had a working base.

3. 🎨 Adding 3D Animations

I integrated:

Rotating 3D objects
Interactive camera movement
Smooth transitions

This gave the site a premium feel.

4. ✨ Enhancing UI/UX

Clean typography
Subtle hover effects
Smooth scrolling
Minimalist design

5. 🚀 Deployment

I deployed the site using GitHub Pages:

Pushed code to a repo
Enabled Pages
Got a live URL instantly

⏱️ Time Breakdown

Task	Time
Planning	5 min
AI Code Generation	20 min
Styling & Animations	20 min
Deployment	5 min
Total	~50 min

🎯 Key Takeaways

AI can drastically speed up development
You don’t need to be a 3D expert to build cool visuals
Combining AI + creativity = powerful results

📈 What’s Next?

Add more interactive 3D elements
Improve performance
Integrate backend for contact form

🙌 Final Thoughts

This project showed me how powerful AI tools like Gemini can be.

What used to take days can now be done in under an hour.

💥 The future of development is AI-accelerated.

🔗 Check It Out

👉 https://sripadh-sujith.github.io/portfolio/

💬 Feedback?

I’d love to hear your thoughts and suggestions!

Feel free to fork, star, or reach out 🚀

New updates on 🔮Celestia an ai horoscope expert

Sripadh Sujith — Sun, 29 Mar 2026 03:18:15 +0000

I have created a ai horoscope expert earlier this month and I have added new updates to it for providing professional horoscope insights with simple prompts
Check it out now: https://huggingface.co/spaces/Sripadh2009/Celestia/