FastAPI & MongoDB Atlas Full-Stack App

Just published a new video in my Web Development Series 🚀 This time, I covered Mongoose Schemas & Models in a very simple and beginner-friendly way. If you're learning Node.js + MongoDB, this is a must-know concept because it helps you structure and validate your data properly. In this video, I explained: • What is a Schema • What is a Model • How they work together • How to create them in Mongoose Perfect for beginners starting backend development. Check it out here 👇 https://lnkd.in/gxEdVnDW #nodejs #mongodb #mongoose #backenddevelopment #webdevelopment #javascript #learnprogramming #jdcodebase

🚀 Learning System Design by Building a URL Shortener Today I worked on building a full-stack URL Shortener to understand how real-world systems are designed and scaled. What the backend is doing: Accepts a long URL and generates a unique short code Stores the mapping in MongoDB Redirects users from short URL → original URL Tracks total clicks for each shortened link Exposes a simple analytics endpoint for URL stats Handles duplicate URLs and validates input properly Tech Stack: Frontend: Next.js, Tailwind CSS Backend: Node.js, Express.js Database: MongoDB Atlas Deployment: Vercel + Render This project helped me understand concepts like: API design Routing and controllers Database schema design URL mapping logic Analytics tracking Deployment and production debugging 🔗 Live Demo: https://lnkd.in/g_qd-WB7 💻 Source Code: https://lnkd.in/g7Kp865n Still learning and improving it step by step ⚡ Would love to hear what features you’d add next! #SystemDesign #FullStackDevelopment #WebDevelopment #Nodejs #Nextjs #MongoDB #SoftwareEngineer #BuildInPublic #BackendDevelopment #100DaysOfCode

Added MongoDB to my tech stack! 🛠️ I'm excited to share my latest backend project — a clean FastAPI + MongoDB Todo API built with best practices. Project Highlights: Full async CRUD operations using Motor (MongoDB's official async driver) Clean architecture with Repository + Service layers Proper Pydantic v2 models for request/response validation Filtering support (?done=true/false) Health check that verifies real MongoDB connection Docker Compose setup for easy local development Well-structured codebase with proper error handling and lifespan events This project was my first deep dive into MongoDB in a FastAPI environment, and I really enjoyed working with its flexible schema design, powerful querying, and high performance for document-based data. Updated Tech Stack: FastAPI + Uvicorn MongoDB + Motor ← New! Pydantic v2 Async Python Repository Pattern + Clean Architecture I'm now confidently applying these patterns to my bigger project Bookie — a social platform for book lovers with microservices, gRPC, Qdrant vector search, and RAG-based recommendations. Full repository (with Postman collection + Docker setup): https://lnkd.in/dGcDz2dm MongoDB has quickly become one of my favorite databases for rapid development and scalable APIs. Excited to keep building with it! Open to feedback and discussions around FastAPI + NoSQL patterns. #FastAPI #MongoDB #Motor #Python #BackendDevelopment #NoSQL #CleanArchitecture #AsyncPython #SideProject

🚀 From Idea to API: Building a FastAPI CRUD App with Neon (PostgreSQL) I recently built a backend CRUD application using FastAPI connected to a cloud PostgreSQL database (Neon). This project helped me understand how real-world backend systems handle data efficiently. 🔹 What I Built A complete Student Management API where users can: • Create student records • Read all / single student data • Update student information • Delete records 🔹 Backend Stack • FastAPI (high-performance Python framework) • SQLAlchemy (ORM for database handling) • PostgreSQL (Neon cloud database) 🔹 Key Learnings • Structuring scalable backend architecture • Connecting FastAPI with a cloud database (Neon) • Writing clean CRUD operations using SQLAlchemy • API testing using Swagger Docs 🔹 What’s Next I’m planning to extend this into a full-stack app by integrating it with Next.js and adding authentication (JWT). This is just a small step towards building production-level systems like SaaS platforms 🚀 I’d appreciate your feedback and suggestions! #SMIT #FastAPI #Python #PostgreSQL #BackendDevelopment #WebDevelopment #CRUD #Neon #LearningJourney

Hitesh Choudhary Just watched "Taking Backend to Next Level" by Hitesh Choudhary 🔥 This part of the Chai aur Code backend series is pure gold if you're serious about MongoDB + Mongoose. Key takeaways: • Moving from basic schemas to real-world data modelling • Understanding when to embed vs reference • Designing scalable and performant MongoDB schemas • Avoiding common mistakes that slow down your backend later If you're building full-stack apps (especially MERN), this video helps you think like a professional backend developer — not just someone who can do CRUD. Highly recommended for intermediate developers who want to level up their backend skills. Watch here: https://lnkd.in/ggBCgQ3s What’s one thing you changed in your data modelling after learning advanced Mongoose concepts? Share in comments 👇 #NodeJS #MongoDB #Mongoose #BackendDevelopment #MERN #FullStack #WebDevelopment

Days 6, 7 & 8 of backend learning from the Masterji backend challenge. In these sessions, I focused on understanding how a proper backend is structured. I learned to organize code into controllers, routes, models, middlewares, and utils, which gave me clarity on how scalable applications are actually built. I also worked on connecting MongoDB Atlas using Mongoose and followed best practices like keeping the DB connection separate, using async/await, and handling errors with try-catch. As sir mentioned, since the database doesn’t run locally like our app, handling async operations properly becomes important. I also understood how environment variables help in keeping configurations secure. Along with this, I explored schema design in more depth — defining fields, adding validations, and structuring data for real-world use cases like ecommerce and hospital systems. Learned how arrays and nested structures help manage more complex data. Overall, these days helped me connect backend structure, data modeling, and database integration into one complete flow. #Nodejs #MongoDB #Mongoose #Backend #FullStack #LearningInPublic #ChaiCode

"It’s just a proof of concept." 🚩 Famous last words in backend engineering. Two years into operating a SaaS LMS built on Laravel, our "prototype" hit a wall. As we decoupled for mobile APIs, we triggered a massive network bottleneck. Our PHP workers were dragging 500k+ rows of raw PostgreSQL data across the network just to compute simple monthly aggregations and user streaks. The Result? Extreme memory exhaustion, lock contention, and 4-second dashboard load times that frustrated our users. The Solution: Respecting Data Gravity. 🌎 Instead of moving the data to the computation, we pushed the computation into the PostgreSQL execution engine: 🚀 N+1 Aggregations → PL/pgSQL Cursors (54% faster) 🚀 Bulk Ingestion → Statement-Level Triggers (88% faster) 🚀 Live Analytics → Materialized Views (99.9% faster) 🚀 Retention Logic → Recursive CTEs (95% RAM reduction) I’ve documented the entire architectural post-mortem, including the raw EXPLAIN ANALYZE outputs and a reproducible Docker environment. (Links to Hashnode Deep-Dive, Kaggle Playground, and GitHub Source in the first comment below! 👇) How do you handle the "ORM Wall" in your stack? Have you experimented with pushing logic into the database engine? Let's discuss. #DataEngineering #PostgreSQL #Laravel #SystemDesign #SQL #BackendPerformance #SoftwareArchitecture #SoftwareEngineering #DatabaseOptimization

Laravel 13 ships with native vector search via whereVectorSimilarTo(). But before you can use it, you need PostgreSQL + pgvector set up on your machine. We published a hands-on tutorial covering everything from installation to querying, all from the terminal on Ubuntu 25.04: 1. Install PostgreSQL 17 and pgvector. 2. Enable the vector extension in a database. 3. Create tables with vector columns. 4. Run similarity searches with 3 distance operators (L2, cosine, inner product). 5. Build HNSW and IVFFlat indexes for performance. 6. Practical example: finding similar products with vector + SQL filters combined. If you are planning to build semantic search or AI-powered features with Laravel 13, this tutorial gives you the database foundation you need first. Read the full tutorial: https://lnkd.in/gnx55uXv #PostgreSQL #pgvector #VectorSearch #Laravel #Laravel13 #Ubuntu #Database #AI #Tutorial

🚀 Day 53–71: The Phase Where Everything Started Making Sense Until now, I was writing code… But in these days, I finally understood how real applications actually work behind the scenes. 👨🏫 Thanks to Siva Kumar Sir at DevGnan for guiding us step by step. 💭 What changed in these days? I moved from just frontend thinking → to understanding the complete flow of an application From clicking a button… to knowing where that data goes, how it travels, and where it gets stored. 🧠 Concepts I Explored ⚡ How backend works using Node.js ⚡ Building servers with Express.js ⚡ Sending responses from backend ⚡ Connecting frontend to backend (real communication) ⚡ Introduction to MongoDB ⚡ Using Mongoose to structure and store data 🔥 What I Built Instead of just learning theory, I implemented a real signup system ✔️ User enters details in frontend ✔️ Data travels through API ✔️ Backend handles it properly ✔️ Database stores it successfully 👉 For the first time, I felt like I built something real, not just practice code. 🔄 The Flow I Now Understand Clearly Frontend → Request → Server → Database → Stored Data Simple line… but powerful understanding 💡 📌 Big Realization Backend is not just code. It’s the brain of the application. And once you understand the flow, everything starts connecting. This phase gave me a lot of confidence. Now I’m not just learning… I’m building with understanding 🚀 #MERNStack #BackendDevelopment #NodeJS #ExpressJS #MongoDB #CodingJourney #FullStackDeveloper

"Ugh, where did I save that link?" I ask myself this every single day. We've all been there. You save something important, think "I'll read this later" and then it disappears into a black hole of browser bookmarks forever. Last month I saved an article about React performance optimization. Three weeks later I needed it. Searched everywhere. Couldn't find it. So instead of just moving on... I built something about it. Introducing LinkVault - an intelligent bookmark manager powered by RAG architecture. 🔗 Live: https://lnkd.in/dNUS4-mF 💻 GitHub: https://lnkd.in/dCt3PqEN Most devs know what RAG is. Very few have actually implemented one end to end. Here's what's actually happening under the hood: → You paste a URL → Cheerio auto-scrapes the title and description → OpenAI converts that text into 1536 numbers (an embedding) → MongoDB Atlas stores those numbers as a vector → When you search, your query also becomes numbers → MongoDB finds the bookmarks whose numbers are closest to yours That's a full RAG pipeline. Built from scratch. No LangChain. No shortcuts. SEARCH FLOW: Type query → query embedded using same model → MongoDB $vectorSearch runs cosine similarity → semantically closest bookmarks returned Why cosine similarity over Euclidean distance? Because embeddings measure direction of meaning, not magnitude. Two texts can mean the same thing with different lengths ; cosine captures that correctly. Why chunk text before embedding? One giant vector for a 3000-word article becomes a blurry average of every topic in it. Chunked embeddings let you match specific sections. Tech: React, Node.js, Express, MongoDB Atlas Vector Search, OpenAI Embeddings, Cheerio, Tailwind, Framer Motion Because "I'll find it later" shouldn't mean "I'll never find it." Let's connect. #BuildInPublic #ReactJS #NodeJS #MongoDB #RAG #VectorSearch #FullStack #SDE