Project Description
I’m building a web-based platform that screens applicants automatically. The core workflow needs to:
• parse incoming résumés/CVs,
• embed both the résumé data and our job descriptions,
• run a retrieval-augmented generation pipeline that compares the two,
• return a clear numerical fit score so the recruiter can decide whether to proceed.
All candidate data will be embedded with FAISS, while LangChain will orchestrate the retrieval and callouts to the OpenAI GPT endpoints. I already have the OpenAI key and hosting in place; what’s missing is the complete back-end logic, the scoring algorithm, and a simple front-end view that lists the ranked applicants and their matching rationale.
Deliverables
1. End-to-end web application (React or similar front end, Python back end) running locally in Docker.
2. Resume parser that extracts structured data (education, skills, experience) ready for vectorisation.
3. RAG module using LangChain + FAISS that produces the fit score and a short explanation paragraph.
4. REST endpoints documented with Swagger/OpenAPI.
5. Brief deployment guide and code walkthrough.
Acceptance criteria: given a sample résumé and job description, the system must return a repeatable score between 0–100 plus an explanation, all under 5 seconds.
If you’ve shipped similar NLP or hiring-tech projects and are comfortable with LangChain, FAISS, and OpenAI, I’d love to see a link or demo in your proposal.