Overview

I am building a mock interview app to help candidates be more confident and prepared for their interviews. I have started building the app but am running into issues and need help debugging.

The flow should go like:

• I upload my resume (Pdf format)
• I input the job description of the role I'd like to prepare for
• I click "Start Mock Interview"
• You parse details for the resume like name, positions they previously held, where did
  they work, and what did they accomplish during those gigs, and dates of times periods of where they worked.
• Use resume details and job description to generate meaningful and personalized questions using OpenAI
• Candidate is brought to a screen with a soundwave of the AI
• AI begins interview using livekit conversational AI
• AI greets candidate and asks to give a quick story of where they started, what has brought them here, and what they are looking for next.
• AI uses generated questions, details from resume, details from job description, and the candidates responses to continue the conversation and ask meaningful questions. -- This goes on for about 15-20 min. Throughout the conversation you should only really ask 5 meaningful questions.
• After the interview is over, the candidate receives a report where we use AI to analyze the transcript and tell the candidate what they did well and what they can improve on. - - If possible, the report should use examples from the interview.

Problem Description

• Candidate can upload resume and input job description, however I am not sure if the resume is being parsed correctly.
• Livekit server room is not getting connected properly
• I am not able to have the Open AI content talk to the liverkit server, and because of this there is no voice, although I am able to collect the content from the candidate's voice.

Acceptance Criteria

• Candidate can upload resume and job description
• Resume gets parsed and details are collected
• Mock Interview begins and livekit server is started
• OpenAI generates questions and is able to constantly go back and further with the livekit server to see what the candidate said most recently
• A candidate is able to talk to an AI and the AI is able to talk to the candidate specifically about this interview. It should be personalized using the resume and job decscription.

Technical Details

• I am using .next, typescript, and python (for the livekit server)

Link to Project

I can email you the files. I will send the following:

• page.tsx
• api/socket/route.ts
• api/token/route.ts
• api/generate-report/route.ts (using to generate a report once the interview ends)
• api/pdf-parse/route.ts (trying to parse pdf)
• api/interview/route.ts (question generation during interview)
• main.py (starts livekit server)

Let me know if you'd need anything else.

Livekit docs: https://docs.livekit.io/home/