WebFeatured
LacakKarirku
2026 Independent product build Solo Software Engineer
An AI-assisted job-search workspace for PDF CV parsing, multi-source job aggregation, match review, application tracking, and follow-up.
Next.jsTypeScriptTailwind CSSBetter AuthNeon PostgresDrizzle ORMGroq AICheerioVercel
What was broken?
Job seekers often save links everywhere, read job descriptions repeatedly, and still do not know which roles are worth applying to first. I wanted one workspace that connects the CV, job source, and next action.
How I fixed it
I built PDF CV upload and parsing, job aggregation with manual entry as a fallback, heuristic CV-to-job matching, LLM-assisted suggestions, and application tracking with follow-up. Neon Postgres and Drizzle store the data, while Better Auth handles authentication.
What it can do
- CV PDF upload and parsing
- Job aggregation with manual input fallback
- Heuristic CV-to-job matching
- LLM-assisted CV suggestions
- Application pipeline tracking and follow-up
The stack that shipped it
Next.js gave the project one place for UI and API routes. Cheerio handled stable sources, Groq handled CV and suggestion tasks, and Drizzle kept the database layer readable.
What changed
A working job-search workspace connecting CV review, job collection, matching, and application progress. Match scores are heuristic guidance, not hiring predictions; external job-source availability can change.
What I learned
The hardest part was not the AI prompt, but making the messy edges of public job sources fail gracefully and still giving users a useful fallback.
