Land Your Dream Job Faster. Upload your resume and get instant, AI-driven feedback to beat the ATS and impress recruiters. Tailored interview prep included.
A Major Project Report submitted in partial fulfilment of the requirements for the degree of BACHELOR OF TECHNOLOGY in COMPUTER SCIENCE & ENGINEERING.
This project was developed by Group No. 23 (CSE) for the academic session 2025 — 2026:
| Name | Roll Number | Department |
|---|---|---|
| Aniruddha Adak | 27600122030 | Computer Science & Engineering |
| Krishanu Banerjee | 27600122136 | Computer Science & Engineering |
| Sayan Chakraborty | 27600122180 | Computer Science & Engineering |
| Soni Kumari | 27600122150 | Computer Science & Engineering |
Under the guidance of:
Mrs. Nivedita Das, Assistant Professor, CSE
Prof. Sagar Chakraborty, Head of the Department of CSE
Institution:
Budge Budge Institute of Technology (BBIT)
Affiliated to MAKAUT & Approved by AICTE, Accredited by NAAC.
The modern job-application process is mediated by Applicant Tracking Systems (ATS) that filter out approximately 75% of resumes before a human recruiter ever sees them. Candidates rarely receive actionable feedback, and existing tools optimize narrowly for keyword density without addressing the broader problem of multimodal input or live interview performance.
CareerZen is a next-generation, open-source career platform that unifies three traditionally separate workflows into a single end-to-end pipeline:
- AI-powered ATS resume analysis
- Voice-based mock interview practice
- A Kanban-style job-application tracker
- Multimodal ATS analysis: Accepts resume as PDF or DOCX, and job description as text or screenshot (OCR). Computes a 0–100 match score with keyword-level optimization tips.
- Voice-based mock interviews: Allows candidates to speak their answers naturally via the browser microphone, transcribing with India-optimized Speech-to-Text, and providing STAR-structured feedback generated by an LLM.
- Persistent application tracker: A Kanban-style board integrated into the dashboard, so that every analyzed resume can be tracked from "wishlist" through "offer" without leaving the platform.
- Open-source transparency: A permissively licensed codebase (MIT) with documented prompts, public Prisma schema, and reproducible benchmarks.
- Frontend: Next.js 16 (App Router), React 19, Tailwind CSS v4, Framer Motion
- Backend: Next.js API Routes (Node.js runtime)
- Database & ORM: PostgreSQL (Neon serverless Postgres), Prisma
- Authentication: Clerk
- AI & Machine Learning: Google Gemini 3.0 Flash (LLM), Gemini Vision (OCR), Sarvam AI (Speech-to-Text)
- Parsing: pdf2json, mammoth
- Deployment: Vercel Edge + Node runtimes
The system has been validated through a combination of unit tests, integration tests, and a user study with fifty sample resumes drawn from the IT domain.
- ATS scoring accuracy: ±4.2% mean absolute deviation against a human-expert baseline.
- Gemini API latency: 2.8 seconds per average analysis round-trip.
- Sarvam STT latency: 1.2 seconds for 30-second audio clips (40-55% faster than generic cloud STT).
- Lighthouse Performance Score: 94 (with perfect 100s on accessibility and best-practices audits).
- License: MIT License
- Repository: github.com/aniruddhaadak80/smart-resume-analyzer
- Live Application: careerzen.vercel.app