AI Career Audit

AI-powered career audit that acts like a brutal FAANG recruiter, analyzing resumes and GitHub profiles to give honest feedback, missing skills, projects, and a 30-day roadmap.

Comment

Inspiration

Students spend months building portfolios, coding on GitHub, and perfecting resumes—but never get real feedback from actual recruiters. We submit hundreds of applications, receive automated rejections, and have zero clue what's wrong. AI Career Audit was born from this frustration: a tool that gives students the brutal, honest feedback they need but rarely get.

What it does

AI Career Audit analyzes your resume (PDF) and GitHub profile like a real FAANG recruiter and delivers: Resume & ATS Scores (0-100 with progress bars) Recruiter Feedback - Brutally honest, actionable insights Roast Mode - Funny but constructive criticism (crowd favorite!) Missing Skills - 5 specific skills you need to learn Suggested Projects - 3 project ideas with tech stacks & difficulty Interview Questions - 3 real questions you'll face 30-Day Roadmap - Week-by-week improvement plan

Just upload your resume + GitHub URL, get instant AI-powered career audit.

How we built it

Tech Stack Frontend: Next.js 16 (App Router), Tailwind CSS, shadcn/ui, Framer Motion AI: Google Gemini API (gemini-3.1-flash-lite for higher quotas) Backend: Next.js API routes, pdf-parse, Octokit (GitHub API) How It Works User uploads resume PDF → text extracted via PDF parser GitHub data fetched using Octokit Both sources sent to Gemini API with structured prompt Gemini returns JSON with all 7 analysis categories Results rendered in interactive dashboard Key Technical Decisions Server-side PDF parsing to avoid ESM/CommonJS issues Retry logic with fallback models to handle quota limits Structured JSON output with responseMimeType for consistent parsing Data cleaning for GitHub null values (description, language)

Challenges we ran into

PDF Parsing Issues: pdf-parse had ESM compatibility problems with Next.js 16. Solution: Used server-side API route with pdf2json. Gemini API Quota Limits: Free tier exhausted during testing. Solution: Switched to gemini-3.1-flash-lite (highest quota model) + added retry logic and fallback models. GitHub Null Values: Repositories returned null for descriptions/languages, breaking Zod validation. Solution: Added data cleaning to replace null with default values. Structured JSON Output: Getting consistent JSON from Gemini required precise prompt engineering. Solution: Used responseMimeType: "application/json" and enforced strict schema in prompts

Accomplishments that we're proud of

Successfully integrated Gemini API for 7 different analysis types in one call Roast Mode - The feature that makes everyone smile and remember our project Beautiful, animated dashboard with progress bars, score cards, and interactive UI Deployed on Google Cloud Run - qualifies for "Best App Deployed on Google Cloud" prize Working prototype in 19 hours - full-stack app from scratch with AI integration Actionable 30-day roadmap - transforms feedback into real progress

What we learned

Gemini API is incredibly powerful - Structured JSON output with specific schemas enables building sophisticated AI applications. Prompt engineering is everything - Detailed prompts with clear examples produce 10x better results. Quota management matters - Knowing which models have higher quotas (Flash-Lite, Flash) is critical for hackathons. Server-side processing is safer - Moving PDF parsing to API routes avoids ESM/CJS conflicts. Error handling makes or breaks UX - Retry logic, fallback models, and clear error messages keep users happy. UI/UX drives engagement - Roast Mode being visually distinct (red card, fire emoji) makes it memorable.

What's next for AI Career Audit

Immediate Plans LinkedIn Profile Integration - Add social/professional network analysis Historical Tracking - Track improvement over time with saved reports Resume Builder - Generate improved resume suggestions based on feedback

Future Features Industry Comparison - Compare skills with industry averages Job Matching - Match skills with specific job requirements Study Resources - Generate learning resources for missing skills Team Mode - Allow peer reviews and collaborative feedback

Enterprise Vision University career centers integration Bootcamp graduation assessments Corporate internal mobility programs

Built With

  • artificial-intelligence
  • career-development
  • docker
  • education
  • framer-motion
  • gemini-ai
  • generative-ai
  • github-analyzer
  • github-api
  • google-cloud
  • google-cloud-run
  • google-gemini-api
  • javascript
  • natural-language-processing
  • next.js
  • octokit
  • pdf-parsing
  • react
  • recruiter
  • resume-analyzer
  • shadcn/ui
  • student-tools
  • tailwind-css
  • web-development
Share this project:

Updates