This project proposes an AI-powered journaling feature for Pearl Discovery, a youth coaching platform. The goal is to help students ages 10–16 identify and name their personal strengths through everyday reflection.
Project Overview
An end-to-end product proposal combining LLM prompt engineering with a full PRD — covering user flows, success metrics, risk mitigation, and a phased pilot plan across student, coach, and parent stakeholder layers.
Prototype
A browser-based proof-of-concept using the Gemini API. Students write a short journal entry and receive a personalized Strength Card mapping their words to one of five pillars: Relationship, Leadership, Curiosity, Resilience, and Integrity. Includes a persistent Strength Log and a coach safety flag system for entries with distress language.
Technical Stack
- Designed structured LLM system prompts to analyze free-text student input and return JSON-formatted strength classifications
- Integrated a human-in-the-loop (HITL) review trigger for flagged entries to ensure COPPA/FERPA compliance and student safety
- Prompt framework validated through worked examples across all five strength pillars
PEARL 