Next Hire
A full-stack AI platform and Extension to automate resume parsing and job application autofill, reducing application time by 40%. Features a fault-tolerant RAG pipeline and hybrid scraping engine.
Timeline
Nov 2025 - Present
Role
Full Stack Developer
Team
Solo
Status
Not CompletedWhat Users Can Do
AI-Powered Resume Parsing: Instantly converts raw PDF into structured data using custom schemas via Gemini 2.0 & OpenAI.
Smart Autofill Extension: Chrome extension that 'reads' job pages and fills complex questions using profile context.
Bring Your Own Key (BYOK): Flexible auth system allowing power users to input their own Gemini/OpenAI keys.
Hybrid Scraping Engine: Polymorphic engine switching between Puppeteer and Cheerio for optimal parsing.
Tech Stack
- Nextjs
- TypeScript
- Tailwind
- PostgreSQL
- Prisma
- Gemini
- Puppeteer
- Zod
Challenges I Faced
Engineered a self-healing JSON parsing layer (jsonrepair) to handle non-deterministic LLM outputs.
Architected a context-aware 'RAG-Lite' engine acting as a retrieval orchestrator.
Built a distributed Token Bucket rate limiter using PostgreSQL transactions for serverless.
Implemented a hybrid polymorphic scraping engine switching between Cheerio and Puppeteer.
How I Overcame Them
Prompt Engineering is Code: Treated prompts as composable, typed software artifacts.
Defensive API Design: Externalized state to DB/Redis for reliable serverless rate limiting.
Stream Handling: Managed long-running AI streams with careful client-side chunk handling.