AI-Driven Generative CAD& Takeoff Desktop App (MVP)

Overview We are looking for a highly skilled, senior-level Full-Stack Developer to build the MVP of IBRIX AI, a cutting-edge desktop application designed to disrupt the architectural, renovation, and construction software market. The application needs to be incredibly fast, visually stunning, and highly secure. The core concept revolves around a powerful, multi-mode CAD data processing and AI generative engine that supports three distinct user workflows: 1. Generative Design from Scratch: The user starts with a blank canvas, inputs specific requirements and constraints via the conversational chat interface (for example: "Design a 70 sqm two-bedroom apartment with an open space kitchen and a balcony, complying with local layout standards"), and the AI engine generates and renders the initial CAD/vector project layout from zero. 2. Conversational Renovation Mode: The user uploads an existing CAD/vector file and types what they want to modify (for example: "Demolish the wall between the kitchen and living room"). The application processes the request, visually renders the proposed modifications on the CAD viewer using standard color-coding (Demolitions in Yellow, New Constructions in Red), and calculates construction costs for the modified parts upon user approval. 3. Direct/Instant Takeoff Mode: The user can simply upload a finished, ready-made project (with no modifications needed). The software automatically parses the entire CAD layout, extracts all quantities, lengths, and areas from the layers, and instantly generates a complete, automated Quantity Takeoff (Computo Metrico Estimativo) based on our price database. We are looking for the best quality at the best price (top-tier value for money). We want an expert who writes clean, professional code without overpricing their services, but we are ready to reward efficiency and technical excellence. Core Architecture and Tech Stack Frontend Framework: Electron or Tauri (Open to recommendations based on performance/bundle size efficiency). Frontend UI/UX: React or Vue.js with modern, clean, minimalist styling. Graphics and Vector Engine: WebGL / Three.js / HTML5 Canvas for high-performance rendering of vector data (parsing, displaying, and dynamically generating .dxf, .dwg, or custom vector formats). Backend Core: Python (FastAPI or Django) running as a secure sidecar/local server. AI Integration: OpenAI API and Anthropic Claude API for LLM-driven spatial reasoning, generative architectural layout processing, and structured geometry output in JSON format. Payment and Licensing: Stripe Integration for subscriptions, user authentication, and seat management. CAD Interoperability: Experience with open-source CAD libraries (like ezdxf) to extract and programmatically generate entity data. Job Requirements and Technical Specifications 1. Desktop App Architecture: Setup the secure communication bridge (IPC or local WebSockets/REST) between the frontend and the Python backend sidecar. 2. Multi-Mode Layout and Cost Engine: Implement a flexible pipeline capable of generating new geometric layouts from text prompts (Blank Canvas Mode), calculating costs on-demand for AI-generated modifications (Yellow/Red logic), OR automatically parsing an unaltered CAD drawing to generate an instant, full project cost estimation. 3. Interactive AI Chat and Generative Parsing: Implement the backend logic to handle conversational text instructions, send requests to Claude/OpenAI, and parse the AI's structural response into actual visual vector geometries (both for creating new layouts and modifying existing ones). 4. Dynamic Vector Rendering: Build an efficient Canvas or WebGL-based viewer capable of loading complex layers, rendering brand-new layouts, and dynamically color-coding modifications based on the AI's proposals. 5. Bulletproof Anti-Piracy and Security: Implement a strict user authentication flow. The local backend must communicate with our licensing server using signed JWT tokens to prevent reverse-engineering. 6. Stripe Bridge: Connect the app licensing to Stripe billing events to lock/unlock features based on active subscriptions. Project Roadmap and Milestones (4 Real Sprints) What we have right now (Assets Ready): Complete UI/UX design, visual branding, and front-end graphics. Fully defined pricing tables and tier structures ready for Stripe mapping. Predefined cost/item databases (tabelle prezzi) to be utilized for the cost calculations. Clear functional specifications of the AI prompt architecture. The Sprints: Step 1: Core Architecture, Auth and Security Bridge Initialize the monorepo, setup the wrapper, embed the Python sidecar, and build the secure JWT authentication/anti-piracy handshake with our licensing server. Step 2: CAD Parsing, Full Drawing Auto-Takeoff and WebGL Viewer Develop the Python module to parse vector data. Build the direct takeoff engine that processes complete, finished drawings to extract total areas/lengths. Build the frontend (WebGL/Canvas) to render drawings and layout grids seamlessly. Step 3: Generative Design, AI Renovation Prompting and Yellow/Red Rendering Implement the conversational AI pipeline for both generating new layouts from scratch (text to blank canvas) and performing renovations (text prompt to AI proposal to visual rendering of yellow/red layers). Connect Stripe for subscription management. Step 4: Cross-Platform Bundling, Testing and Production Release Bundle the complete application into native, optimized installers for Windows (.exe) and macOS (.dmg), including the embedded Python environment. Run performance and security stress tests. Post-Launch Warranty and Support (Crucial Requirement) Included Support Period: We require at least 3 months of post-launch assistance and warranty included in the project scope. Bug Fixing: This period must cover the free fixing of any bugs, glitches, memory leaks, or malfunctions discovered after deployment. How to Apply To ensure you have read the full description and are not using an automated bot, please start your proposal with the word "IBRIX". Applications without this word at the very beginning will be automatically ignored. Please submit your proposal along with relevant portfolio examples of desktop apps you’ve built or WebGL/vector processing tools you have engineered.