HR Fashion — Agentic AI Retail & Careers Platform

Building the AI chat assistant to answer from real, live product/job/service data instead of hallucinating required implementing genuine function/tool-calling with the Groq API rather than a static prompt — the assistant had to query the actual database through defined tools mid-conversation. The AI interview system added another layer: reliably parsing varied resume formats (PDF/DOC) with unpdf, then generating role-specific questions that combined resume content with job requirements, and scoring free-text answers consistently on a 0-100 scale with structured feedback.

Add a non-AI fallback path for the interview scoring system similar to the one already built for recommendations, to keep the feature reliable if the AI provider has downtime. Expand the recommendation engine to factor in review/rating behavior alongside wishlist activity for sharper personalization.