FairHire — AI-Powered Recruitment Automation System (AISIP Capstone Project)
Text
Prompt + Response Writing (SFT)Built a bias-aware AI recruitment screening system by engineering system prompts and evaluating LLM outputs for fairness and reliability. Parsed free-form LLM responses into structured ATS-ready fields and implemented error handling for consistent pipeline execution. Tested the prompt against multiple candidate profiles to check bias effects and verified the system’s output quality and scoring behavior. • Designed an n8n workflow that screens applicants using a Groq LLM (Llama 3.3 70B) • Wrote a bias-elimination prompt that instructs the model to ignore name, nationality, university prestige, and CV formatting • Evaluated outputs and performed a bias-elimination test across candidate profiles • Converted AI plain-text results into structured fields (Score, Qualified, Strengths, Gaps, Summary) feeding Google Sheets analytics