INTERVIEW SCHEDULER
AI-Powered Availability Matching & Calendar System
The Problem Statement
“Coordinating interview availability between recruiters and candidates involves multiple back-and-forth emails, manual calendar checks, and frequent scheduling conflicts. Most scheduling tools force structured time-picker inputs, failing when users want to express availability naturally. This scheduler solves the coordination problem by accepting free-text availability from both sides, using GPT-4 to parse natural language into structured time slots, and running a matching algorithm to find the optimal interview window — then automating the confirmation with .ics calendar invites.”
The Architecture Layout
The backend is a Node.js Express server that exposes availability submission, matching, and confirmation endpoints backed by MongoDB with Mongoose ODM. When a user submits free-text availability, the server sends the input to OpenAI's GPT-4 API with a structured prompt and parses the JSON response into a normalised time-slot array. The matching engine performs set-intersection logic across interviewer and candidate availability arrays, scoring candidate slots by optimality. On confirmation, the server uses SendGrid to dispatch professional HTML emails with dynamically generated .ics calendar attachments compatible with Google Calendar, Outlook, and Apple Calendar. The Next.js 14 App Router frontend provides both a structured time-picker and a free-text input path with React Hook Form validation.
Architecture Design Diagram
