The problem
Job seekers often send CVs that applicant tracking systems (ATS) parse badly, and they rarely get to practise interviews with useful feedback. This platform tackles both: it helps candidates build ATS-friendly CVs tailored to a job description, then lets them rehearse with a voice-based mock interview.
My role
Full-stack development of the platform.
Stack
- Next.js
- TypeScript
- PostgreSQL
- Prisma
- Groq LLM
- Whisper
Architecture
A Next.js and TypeScript app handles the UI and server logic. Resume generation, parsing and job-description optimisation call a Groq-hosted LLM. In the mock interview, Whisper turns the candidate's spoken answers into text, the LLM generates feedback, and text-to-speech reads questions and feedback aloud. Data is stored in PostgreSQL through Prisma.
Key technical decisions
- Groq LLM for fast responses, so resume suggestions and interview feedback feel immediate.
- Whisper speech-to-text so candidates answer out loud, as they would in a real interview.
- Job-description optimisation that compares a CV against a target role before generating changes.
- PostgreSQL and Prisma for typed access to users, resumes and interview sessions.
Result
Candidates can generate an ATS-friendly CV, optimise it for a specific job description, and get instant feedback from a voice-based mock interview.
Links
There is no public link for this project yet. I am happy to walk through it in an interview.