An AI-powered recruitment assistant that uses Retrieval-Augmented Generation (RAG) to match candidates with job openings and provide natural-language insights to recruiters.
This system helps recruiters screen resumes faster by combining local LLM inference with RAG-based retrieval over resume and job data. Recruiters can query candidate information using natural language instead of manually filtering spreadsheets.
- AI-powered resume analysis — suggests suitable job matches for candidates based on skills and profile
- RAG-based retrieval over resume and job posting data for accurate, context-aware responses
- Natural language queries — recruiters can ask things like "show me top-ranked candidates" or "who applied most recently"
- Automated candidate profiling with Firebase integration
- Fast response time — ~5-8 seconds per query
- LLM: Ollama (phi3:mini)
- Retrieval: RAG (Retrieval-Augmented Generation)
- Database: Firebase
- Frontend: Flutter (Dart) — cross-platform (Android, iOS, Web, Windows, macOS)
- Resumes and job data are ingested and indexed for retrieval
- Recruiter submits a natural language query (e.g. "shortlisted candidates for backend role")
- RAG pipeline retrieves the most relevant resume/job chunks
- Ollama (phi3:mini) generates a natural language response using retrieved context
- Results are returned to the recruiter within 5-8 seconds
- Retrieve top-ranked candidates for a specific role
- View shortlisted or most recent applicants
- Get AI-suggested job matches based on candidate skills
Developed as a personal project (March 2025 - August 2025) to explore practical applications of LLMs and RAG in real-world recruitment workflows.