TalentFind

• 

  Website's icon

Inspiration

Recruiters today face an overwhelming challenge: hundreds to thousands of applications, disorganized emails, and repetitive manual filtering that wastes hours before even reaching a usable shortlist. At the same time, students and recent CS/tech graduates feel the system is unfair — they apply everywhere, rarely get seen, and often get filtered out simply because their applications get buried.

We wanted to fix both pains at once:

• Recruiters should instantly see the right candidates, not all candidates.
• Students should finally get a fair, skill-based, AI-driven evaluation that goes beyond a one-page resume.

TalentFind exists to bridge this gap with real AI portfolio analysis and personalized candidate ranking.

What it does

TalentFind is an AI-powered recruiting platform that:

• Analyzes resumes, portfolios, and GitHub repos using AI.
• Reads job descriptions and evaluates candidate fit.
• Automatically ranks and filters the top-matching candidates for each job posting.
• Provides students with an AI Review that inspects their portfolio websites + repos and returns up to five actionable improvements they can make.
• Gives recruiters a streamlined dashboard with shortlist, hire button, and email workflows.

Both sides get clarity:

• Recruiters instantly see who is most qualified.
• Students understand why they weren’t shortlisted and how to improve.

How we built it

1. Initial design + architecture with an external LLM
   We brainstormed and refined the system architecture, workflows, and UI using ChatGPT.

2. Prompt file creation
   A structured prompt file was created containing all platform rules, behaviors, and logic.
   This file was then uploaded into Base44 as context for consistent outputs.

3. Development using Base44 (Claude Opus 4.5)
   Base44 handled:

   • Recruiter dashboard
   • Student dashboard
   • AI portfolio review flow
   • Job posting and candidate filtering
   • AI ranking logic
   • The “Hire” button workflow
   • UI states and transitions

4. AI workflow integration

   • AI reads job descriptions + additional documents
     
   • AI evaluates candidates’ GitHub repos and portfolio content
     
   • AI produces a ranked shortlist
     
   • Students get detailed improvement suggestions
     
   • Recruiters get simple, actionable outputs

Challenges we ran into

• Adding new features broke earlier functionality.
  When we introduced automatic email workflows and the Hire button flow, it temporarily broke candidate matching and ranking logic.
• Maintaining consistent AI behavior when expanding the feature set.
• Ensuring meaningful AI portfolio analysis and improvement suggestions.

Accomplishments that we're proud of

• The AI system fully works — it reads job descriptions, evaluates resumes/portfolios, checks GitHub repos, and ranks candidates.
• Students get real improvement feedback — the AI suggests up to five changes based on their portfolio and repos.
• Recruiters see only the best candidates — filtering and ranking save hours.
• The UI and workflows feel polished due to iterative prompt refinement.
• The platform genuinely solves a real-world problem for both recruiters and students.

What we learned

• How powerful structured prompt engineering is when developing with AI.
• The importance of building a stable prompt file before scaling features.
• How to maintain consistent behavior while adding complex AI-driven functions.
• That AI tools can dramatically accelerate development when guided strategically.
• Building real products with AI is not only possible — it's fast and effective when done right.

Built With

• base44