Resumix AI — Private Candidate Screening Platform
AI candidate screening platform built to analyze resumes, reduce screening latency, and rank candidates using Next.js 14, Express, Redis, and LLM APIs.
| TechnologyTeknologi | Role in SystemPeran dalam Sistem |
|---|---|
| Next.js 14 | Fullstack Framework |
| Express/Redis Queue | Async Task Queue |
| Vector DB | Semantic Search Index |
| LLM APIs | LLM Inference Engine |
Manual HR screening latency (5-10 mins/CV), high LLM inference costs for scanned PDFs, PII bias, and opaque ATS scoring.
Built Resumix AI with Next.js 14 Web Portal, Express BFF Gateway, Redis Task Queue, FastAPI Python microservice with zero-text PDF short-circuiting, structured LLM PII-safe parsing, vector database embeddings, and explainable scoring (45% Cosine Match, 30% Mandatory Skills, 20% Experience, 5% Bonus).
Resumix AI is engineered as a candidate screening platform combining a decoupled microservices architecture, asynchronous task queues, structured LLM parsing, vector search similarity matching, and an explainable multi-factor scoring algorithm.