How maxim-kondratjev/bootcamp_ds_2025 works
Project type and purpose
• This is a full‑stack AI‑powered web application (“Interviewee Agent”) built with FastAPI (Python) on the backend and plain HTML pages on the frontend.
• It exposes REST endpoints for conducting and evaluating mock interviews, using OpenAI Agents under the hood with custom prompts and tools, plus speech‑to‑text (STT), text‑to‑speech (TTS), and text‑to‑text (TTT) modules.
File structure
• Top‑level:
• .github/workflows/ – CI/CD pipeline deploying to DigitalOcean.
• app/ – main Python package.
• config/ – Nginx + FastAPI config for production.
• requirements.txt, setup.sh, dev.sh, .env.example – setup & runtime scripts.
• app/ contains:
• api/ – FastAPI routers (interview.py, evaluation.py).
• agents/ – two agent classes (interviewee_agent.py, evaluation_agent.py), with prompts/ (YAML templates + utils) and tools/ (action/result extraction).
• model/ – STT, TTS, TTT processing modules.
• core/ – app config, constants, OpenAI client wrappers.
• frontend/ – static HTML pages.
• main.py – FastAPI app instantiation + route mounting.
README for additional insights
• Confirms FastAPI + Uvicorn server, OpenAI Agent SDK usage.
• No persistent database is mentioned; state is transient per HTTP call.
• Deployment uses Nginx as reverse proxy on a DigitalOcean droplet.