Login page
Main page
Review result
Review History page

Inspiration

We wanted to make something useful for students.
A lot of students have resumes, but they are not sure what to fix for a real job post.
So we made an AI Resume Reviewer to give fast and simple feedback.

What it does

Users can sign up, upload a resume (PDF/DOCX/TXT/MD), and paste a job description.
Then the app gives review result which include overall score, strengths, weaknesses, things to improve, better bullet examples and missing keywords The app also saves each result, so users can check review history later.

How we built it

Frontend: Next.js + React
Login and database: Firebase Authentication + Firestore
AI feedback: OpenAI API (GPT-5 mini)
File reading: server-side text extraction for PDF/DOCX/TXT/MD

Challenges we ran into

We got a Firebase login error (auth/api-key-not-valid) because our .env.local format was wrong.
OpenAI setup was confusing at first because ChatGPT subscription and OpenAI API billing are separate.
We got a model error when using GPT-5 mini with an unsupported parameter (temperature), so we removed that setting.

Accomplishments that we're proud of

Built a full working app with login, resume upload, AI review, and history.
Added support for multiple file types (PDF, DOCX, TXT, MD).
Connected AI feedback to a clean UI and saved results for each user in Firestore.
Added fallback logic so the app can still run even without an OpenAI key.

What we learned

How to connect frontend, auth, database, and AI API in one project.
How to debug real setup issues with env variables and API keys.
How to build a better user flow for upload, review, and saved history.
How to make feedback clear and useful instead of only showing raw model output.

What's next for AI Resume Reviewer

Add clearer score breakdown (format score, keyword score, impact score).
Add export options like PDF summary.
Improve keyword matching and ranking.
Try a small custom model later and compare it with the current API-based version.