Skip to content

Perplexa-AI Mini is a personal AI assistant inspired by Perplexity.ai, built in the GenAI With JS Batch. Version-2 adds Clerk authentication, stronger memory for LLMs, and an upgraded multi-agent system.

Notifications You must be signed in to change notification settings

YashPandey1405/Perplexa-AI

Repository files navigation

Perplexa-AI Mini | GenAI With JS Batch Journey

🚀 About Me

Hi! I’m Yash Pandey, a Full-stack MERN Developer and Data Scientist with deep expertise in JavaScript, TypeScript, Python, and Java.

  • Proficient in frameworks like Next.js, React, Node.js, Express.js, and Material-UI.
  • Skilled in database design & management: MongoDB, PostgreSQL, MySQL, Redis, and Prisma ORM.
  • Experienced in building scalable backend systems with JWT authentication, WebSockets, Redis caching, and CI/CD pipelines.
  • Strong foundation in Data Science & AI, including Machine Learning, Deep Learning, and Data Visualization (NumPy, Pandas, TensorFlow, PyTorch).
  • Solved 700+ DSA problems, led backend development in hackathons, and participated in competitive coding events.
  • Recent focus: Generative AI integration, OpenAI SDKs, Agent architectures, and LangGraph alternatives.
  • Hands-on with AWS deployment (EC2, Elastic Beanstalk) and Docker containerization.
  • Passionate about building real-time collaborative systems and optimizing full-stack performance for large-scale applications.

🧠 About Perplexa-AI Mini

Perplexa-AI Mini is a personal AI assistant project, inspired by Perplexity.ai, built during my GenAI With JS Batch.

Features

  • Text, Image & Voice Queries → Handles multiple input types.
  • Dynamic AI Responses → Powered by OpenAI LLMs.
  • Real-time Interaction → Smooth user experience with instant responses.
  • Mini Knowledge Base → Stores chat history and context.
  • Authentication & Persistence → Secured via Clerk and MongoDB.

Tech Stack

  • Frontend: Next.js, React, Tailwind CSS, Framer Motion
  • Backend: Node.js, Express.js, REST API
  • Database: MongoDB
  • Authentication: Clerk
  • AI & NLP: OpenAI LLMs
  • Deployment & CI/CD: Docker, AWS EC2

Project Screenshots

Perplexa-AI Banner

Perplexa-AI Banner


📚 GenAI With JS Batch Journey

I completed a structured Generative AI Batch with Next.js projects for hands-on experience.

Highlights

  • 12 Main Lectures + Bonus Videos
  • Topics included: GPT Transformers, Prompt Engineering, Agentic AI, RAG, LangGraph, OpenAI Agent SDK, LLM Memory Management, Graph Databases, Conversational AI, Model Context Protocol
  • Bonus: Docker deployment, Fine-Tuning via Ollama

Key Projects

Lecture Project / Repo Date Key Learnings
1 Tokenizer Visualizer 29th Aug 2025 GPT Transformers, Tokenization, Vector Embeddings
2 System Prompts GenAI 2nd Sep 2025 Prompting, GIGO Principle, System-Prompt Engineering
3 AgenticAI 6th Sep 2025 Agentic AI, COT-Prompting, Multi-Agent Mini-Project
4 RAGify NextJS 8th Sep 2025 RAG Basics, Vector DB Integration, LangChain Importance
5 N/A 10th Sep 2025 Advanced RAG System Design, HyDE Embeddings
6 LangGraph BranchFlow 13th Sep 2025 LangGraph Overview, Nodes & Edges, Workflow Tracing
7 Multi-Agent Playground 15th Sep 2025 OpenAI Agent SDK, Agents & Tools, Workflow Management
8 Safe Bank Assistant 16th Sep 2025 I/P & O/P Guardrails, Human-in-the-loop Agents
9 Mem0 Demo 17th Sep 2025 LLM Memory Management, Short & Long Term Memory
10 Neo4j Integration 18th Sep 2025 Graph Databases, CYPHER Commands, Mem0 + Qdrant Integration
11 Yash AI Voice Assistant 19th Sep 2025 Conversational AI, Realtime Agent, WebRTC with NextJS
12 N/A 20th Sep 2025 Model Context Protocol, MCP Architecture, RAG & Agentic AI
13 N/A 22th Sep 2025 Docker Deployment, Image Optimization, Ollama Fine-Tuning

For detailed notes, check My Notion Notes.


🔗 Links


Made with ❤️ by Yash Pandey

About

Perplexa-AI Mini is a personal AI assistant inspired by Perplexity.ai, built in the GenAI With JS Batch. Version-2 adds Clerk authentication, stronger memory for LLMs, and an upgraded multi-agent system.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages