JobsYo: Building an AI-based but Human-driven Job Search, Research and Apply Ecosystem

Live demo of JobsYo, an AI-powered job search platform I built to automate & organize job hunt - built based on my own experience. I’ll show the technical internals:

1. Multi-Model AI Chatbot dedicated for Customized Application and Connection: How I built a unified abstraction layer that routes between GPT-5, Gemini 3, and Claude based on user preferences, with encrypted API key storage and streaming response handling across all providers.

2. Context Engineering: The unified prompt system that dynamically injects user profiles, parsed job requirements, base resumes, and linked networking connections into a single context window - enabling AI to generate highly tailored responses.

3. Agentic Job Discovery: Orchestration function that coordinates 4 different job APIs (Adzuna, JSearch, RemoteOK, custom scrapers), handles deduplication, and generates AI recommendations for failed sources.

4. Profile-to-Job Matching: AI-powered analysis that compares parsed user profiles against job requirements and provides actionable improvement suggestions.

All running on React + Supabase Edge Functions with real-time streaming.