Reproduce “Zara: An LLM-based Candidate Interview Feedback System” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Nima Yazdani, Aruj Mahajan, Ali Ansari
An LLM-driven recruitment support system, Zara, provides personalized and structured post-interview feedback to candidates and automates query resolution using OpenAI's GPT-4o. This system successfully resolved 75% of candidate inquiries and achieved a 4.37/5 NPS for AI interviews, alongside improved technical question quality and conversational dynamics.
Key insights
- Large Language Models can be effectively deployed to provide scalable, personalized candidate support, addressing a long-standing challenge in recruitment feedback.
- Implementing chain-of-thought and few-shot prompting with explicit guardrails enables LLMs to generate structured, constructive, and role-specific feedback while mitigating negative language.
- Integrating RAG systems for query resolution substantially reduces the administrative burden on recruiters and ensures timely, consistent candidate communication.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.