FastAPI for AI Projects: Efficient Backend Development

🚀 Built an AI-powered dev tool — DevPilot (MVP) I wanted to understand how tools like Copilot/Cursor work, so I built one. 🔧 DevPilot can: Index a codebase Store embeddings in a vector DB Let you ask questions about your code Example: “Where is bubble sort implemented?” → finds the file + explains it using your own code. 🧠 Stack: Python, FastAPI, Qdrant, Ollama (local LLM), CLI 📌 Next (Phase 2): Better retrieval (hybrid search) Function-level chunking VS Code extension Great learning on RAG, backend design, and AI systems. Would love feedback 🙌 #AI #Python #Backend #RAG #DevTools

Developed a full-stack AI-powered coffee shop application featuring a RAG-based chatbot capable of answering questions, taking orders, and providing recommendations. Built using React, Express.js, Python FastAPI, and OpenAI GPT-4o-mini. View the live site: https://lnkd.in/g3wGug_m Demo account: Email: shinji@nerv.com Password: 123456 The site may take a moment to load because it spins down on inactivity. View the project on GitHub: https://lnkd.in/gP6MGg4a

🚀 Excited to share my project: BookSage AI – Hybrid Book Recommendation System! 📚🤖 I built BookSage AI, a machine learning powered system that recommends books using a Hybrid Recommendation Model combining: 🔹 Collaborative Filtering (KNN) 🔹 Content-Based Filtering (TF-IDF + Cosine Similarity) The system is designed with a FastAPI backend and React frontend, providing fast and personalized book recommendations. 💡 Tech Stack: Python | FastAPI | React | Tailwind CSS | Scikit-learn | Pandas | NumPy | Docker 🎥 Watch the Demo in the video below 🌐 Live Demo: https://lnkd.in/gcuSi_6b 💻 GitHub Repository: https://lnkd.in/gmS3cVTq I would love to hear your feedback and suggestions! 😊 #MachineLearning #AI #RecommendationSystem #Python #React #FastAPI #DataScience #Projects #OpenSource #GitHub

Just built and deployed a fully functional AI-powered blog generation tool that helps users create SEO-optimized blogs in seconds. It is completely free as i used the grok api key. Here's what I built: • AI-powered SEO blog generator (Groq + Llama 3.3 70B) • Full markdown editor with live preview • Blog library to save & manage posts • Download blogs as .md or .txt • REST API backend with FastAPI + Python • Dark & Light mode for better user experience • Groq API (free tier) for AI generation • No database needed browser storage 🔗 Live: https://lnkd.in/d5Ff4acb #Python #FastAPI #AI #WebDevelopment #LLM #OpenSource #BuildInPublic

I just published my first open-source project on GitHub. 🚀 It's a lightweight Python router that lets you call GPT-4o, Gemini, Claude, and Mistral through a single function. Swap models with one parameter change — no refactoring. 🔁 As a PM, I've always believed you make better product decisions when you understand the tech underneath. 🧠 Still learning. But shipping things is how I learn best. 🛠️ https://lnkd.in/eEKaiWN2 #AI #ProductManagement #GenerativeAI #LearningInPublic

🚀 Excited to share my Multi-Document RAG Chatbot built as a Python Full Stack project! This system lets users upload PDFs and ask questions in real-time, powered by: Backend: FastAPI + ChromaDB + token-aware chunking Frontend: React + streaming responses AI: Groq LLM for intelligent, context-aware answers It demonstrates full stack development, vector search, and GenAI integration — perfect for real-world AI applications. Check out the demo 👇 and DM me for the GitHub repo if you want to explore the full code! #Python #FullStackDeveloper #FastAPI #ReactJS #RAG #VectorSearch #LLM #GenAI #AIApplications #ChromaDB

Built a Semantic Search Engine from scratch! 🚀 Just wrapped up a project that sounded complex but turned out to be surprisingly achievable. What it does: Scrapes Wikipedia content Splits text into meaningful chunks using semantic similarity Generates vector embeddings using all-MiniLM-L6-v2 Stores them locally as JSON FastAPI backend with cosine similarity search Clean frontend for interaction Tech Stack: Python • FastAPI • Sentence Transformers NumPy • JSON • HTML/CSS/JS The magic? You can search with natural language like "Who influenced his music?" and it understands the meaning, not just keywords. Key takeaway: Building a search engine in 2024 is actually accessible. The hard part isn't the tech anymore - it's the architecture and UX. Would love to hear your thoughts! What should I search next? 🔍 #SemanticSearch #FastAPI #MachineLearning #Python #WebDev #AI #BuildInPublic

Just completed testing of our rebuilt Signal pipeline. The new Python + Firecrawl system is working beautifully - finding high-quality articles and generating content automatically. Big upgrade from the failing Node.js system. Sometimes you need to rebuild from the ground up. #AI #Automation #TechLeadership

Nine minutes. That is how long it takes to go from a blank VS Code window to a working Scrapy spider with Web Scraping Copilot. We built it because generic AI coding assistants are good at code and bad at scraping. They do not know Scrapy project structure. They skip page objects, generate selectors against the full DOM and those selectors break. They do not write fixtures. They give you a single file that works until it does not. Web Scraping Copilot fixes this. Read our announcement blog https://lnkd.in/dvceQtvf It is a free VS Code extension that runs a local MCP server, brings Zyte's scraping expertise into whatever coding assistant you already use, and generates a proper Scrapy project: spider, page objects, items, fixtures. 👉 Free to use. Try it out and let us know what you think: https://lnkd.in/dgmCWQgd #WebScrapingCopilot #Scrapy #Python #WebScraping #VSCode #DataEngineering

🧠 I built an AI-Powered Quiz Master using Python and Google Gemini! I wanted to create a tool that makes learning any subject interactive and personalized. So, I developed a desktop application that generates custom MCQs on the fly using Generative AI. Key Features: Gemini 3 Integration: Uses AI to generate context-aware questions in real-time. Adaptive Difficulty: The app tracks your score and adjusts the next quiz level (Easy/Medium/Hard) automatically. Modern UI: Built with CustomTkinter for a sleek, dark-mode experience. Multi-threading: Ensures the UI stays responsive while the AI is "thinking." The Tech Stack: 🐍 Python 🤖 Google Generative AI SDK 🎨 CustomTkinter (GUI) 🧵 Threading & JSON parsing Check out the demo below! I'm looking forward to expanding this with features like PDF exports and a study timer. #Python #ArtificialIntelligence #GoogleGemini #SoftwareDevelopment #CodingProject #GUI #BuildInPublic