AI Native Job Board

Inspiration

Hiring developers at my company is slow and noisy. Resume screening takes a lot of time and still doesn’t reliably reflect real coding ability. I wanted to explore whether hiring decisions could be based on actual proof of work instead of resumes and keywords.

What it does

AI-Native Job Board replaces resume screening with artifact-based evaluation. Recruiters create task-focused job descriptions, and candidates apply by submitting GitHub commits or repositories. AI evaluates the code against the job requirements and produces a relevance score and explanation, helping recruiters shortlist candidates faster and more objectively.

How I built it

I built the MVP using Next.js, with AI-assisted job description rewriting and code evaluation powered by Gemini 3. The system fetches GitHub artifacts, analyzes code structure and patterns, compares them against task-level job descriptions, and surfaces ranked results in a recruiter-facing leaderboard.

Challenges I ran into

The biggest challenge was keeping the scope realistic while still demonstrating an AI-native system. Fetching and analyzing code artifacts reliably and making AI outputs explainable rather than opaque were also key challenges.

Accomplishments that I’m proud of

I’m proud of shipping a complete end-to-end flow that solves a real hiring problem I face personally. The system demonstrates a clear shift from resume-based screening to proof-of-work-based evaluation within a tight time constraint.

What I learned

I learned that hiring isn’t a matching problem—it’s a learning problem. Real signals come from work artifacts, not credentials, and AI systems are far more valuable when they assist human decision-making rather than replacing it.

What’s next for AI-Native Job Board

Next, I want to expand beyond developer roles by supporting different artifact types, improve scoring using feedback from real hiring outcomes, and evolve the system into a continuous learning platform for hiring.

Built With

• gemini
• githubapi
• langchain
• nextjs
• supabase