How We Built “Job Interview Coach” with ChatGPT and OpenAI
A one-page app powered by React, Flask, and carefully crafted prompts — built in collaboration with AI from idea to execution.
· 2 min read

We've been exploring the creative boundaries of ChatGPT, and as a fun challenge, we decided to build a single-page web app with a few constraints:
ChatGPT had to come up with the idea and help refine it
ChatGPT had to write the code
The app had to use the OpenAI API to deliver something fun or useful
The result? Job Interview Coach — an AI-powered tool that generates customized interview questions tailored to specific roles and interview stages. ChatGPT even named the app!
👉 Try it here: https://xoself.com/job
The Tech Stack
We built the app using React/Next.js on the frontend, paired with a lightweight Flask/Python backend hosted on Azure. All prompts are processed using the OpenAI API, with dynamic input fields that let users select the job type and stage of the interview.
The heart of the app is this carefully engineered prompt:
prompt = (
f"You are an experienced job interviewer and an expert in the {job_type} industry. You are responsible for creating an interview question for a candidate.\n\n"
f"Job Type: {job_type}\n"
f"Interview Stage: {interview_stage}\n"
"When generating interview questions ensure that:\n"
"- You are an expert in the {job_type} industry.\n"
"- Your questions are highly specific to the {job_type} and include industry-specific terms of art.\n"
"- Your questions reference recent trends and issues within the industry.\n"
"- Your questions are tailored to the {interview_stage} and become more specific based on the stage (phone interview, technical interview, behavioral interview, final interview).\n\n"
"Additionally, provide only an interview question for a {job_type} at the {interview_stage} stage. Do not include any explanations, comments, or context—just the question. Remove any leading quotes or numbers.\n\n")
The prompt ensures specificity, relevance, and increasing difficulty based on the selected stage. The AI returns a sharp, targeted question — no fluff, no filler.
Why This Matters
The AI behind this app doesn’t just generate content — it writes the code, engineers its own prompts, and delivers context-aware output with stunning reliability. Its breadth of domain knowledge is outstanding and it demonstrates that ChatGPT is an incredibly versatile tool for rapid development.
Let us know what you think — and feel free to send feedback or feature suggestions
!
Stay Updated
Subscribe to get more insights delivered to your inbox.
Originally published on Substack.