How zoomish/interview-assistant works
Project type and purpose
• This is a backend microservice: a Telegram bot for interview preparation built with Nest.js.
• It uses the node-telegram-bot-api to interface with Telegram, Postgres via TypeORM for persistence, and the Gemini AI API for dynamic question generation.
File structure
• Top‑level Nest.js app: src/app.module.ts, src/main.ts
• Bot module: src/bot
• bot.module.ts & bot.service.ts – core Telegram integration
• services/handling – command‑level handlers (greeting, help, AI generation, speech‑to‑text, etc.)
• services/callback – inline‑keyboard callbacks (announcement, history, interview, review, user profile edits)
• Domain modules:
• src/user – User entity, DTOs, UserService
• src/history – History entity, DTOs, HistoryService
• src/review – Review entity, DTOs, ReviewService
• Configuration: example.env, nest-cli.json, tsconfig.json
• Dependencies: node-telegram-bot-api, @nestjs/*, TypeORM, pg, Gemini API client
README for additional insights
• Features: profession- and skill‑based Q&A, AI‑generated dynamic content, persistent user profiles & histories
• Deployment hints: long‑polling vs webhook, environment variables for Telegram token, Gemini API key, Postgres credentials