BackCheck

🚀 About the Project

🔍 Inspiration

Hiring is messy. Resumes don’t always tell the full story, and recruiters rarely have time to verify every link or claim. We wanted to change that. The idea for BackCheck came from a simple need:
Can we turn a resume into a real-time, citation-backed intelligence report—instantly?

With Perplexity’s Sonar API, we had the perfect AI research engine to make it happen.

🛠️ What We Built

BackCheck is an AI-powered platform that:

• Accepts a candidate’s PDF resume
  
• Extracts all embedded URLs (LinkedIn, GitHub, portfolios, etc.)
  
• Uses Perplexity Sonar to conduct real-time, trustworthy research
  
• Compiles a human-readable report summarizing skills, credibility, achievements, and red flags
  
• Visualizes skills in a clean dashboard
  
• Includes an interactive chat to ask deeper questions about the candidate

Built with a Next.js + Tailwind frontend and a FastAPI backend, BackCheck delivers deep background insights in seconds.

💡 What We Learned

• Leveraging Perplexity’s Sonar Pro API for deep candidate analysis
  
• Designing smart prompts and chaining logic using LangChain
  
• Skill extraction and grouping from unstructured resume text
  
• Building modern UIs quickly with Tailwind CSS and Recharts
  
• Handling the weirdness of PDF parsing and web scraping with PyMuPDF + BeautifulSoup
  

🧗‍♂️ Challenges We Faced

• Resume parsing is chaotic—PDFs are never consistent
  
• Prompt engineering for consistent Sonar API responses was trickier than expected
  
• Handling rate limits while maintaining a fast user experience
  
• Coordinating frontend, backend, and AI logic under tight hackathon pressure 🧠⚡
  

❤️ Why We’re Proud

We didn’t just ship a prototype—we built something immediately useful.
BackCheck gives hiring teams a superpower: fast, reliable insights about candidates grounded in real research, not guesswork.

It’s AI-powered hiring intelligence—finally done right.

Built With

• next.js
• perplexity
• recharts
• tailwind
• uvicorn
• vercel