03 / PROJECT DETAIL
CareerOS
Job tracker, task manager and AI notes in one app.
Solo project
SUMMARY
Smart notes use semantic search (vector embeddings), plus a Gemini-powered career coach and task planner with model fallback.
PRODUCT
PROBLEM
Built from personal frustration: job applications scattered across browser tabs and notes, no way to track progress through an online course, no deadline reminders, and notes that could only be found by exact keyword.
SOLUTION
One app with four modules: a job tracker, a task manager with categories, priorities and daily progress logs, smart notes with semantic search, and AI tools for career coaching and learning plans.
KEY DECISION
Model fallback chain
Each AI endpoint tries a primary model, falls back to the next model if it fails or hits quota, and finally returns a static response, so the interface never breaks.
IMPLEMENTATION
- Next.js
- TypeScript
- Prisma
- PostgreSQL
- pgvector
- Supabase
- Gemini