Content Directory for Ursa's RAG Knowledge Base

This directory contains markdown content files that will be processed into vector embeddings for Ursa's RAG (Retrieval-Augmented Generation) system.

⚠️ IMPORTANT: Source of Truth

The authoritative source content is located in `/rag` directory at the project root.

• /rag/personal/personal.md - Comprehensive biography, experience, skills, and professional philosophy
• /rag/case_studies/ - Detailed project case studies organized by type:
• product_design/personal/ - Personal projects (Vibio, Aether)
• product_design/industry/ - Industry projects (DriQ Health, Sparto, Synofin, etc.)
• branding/ - Branding projects

Content in `_content/` is DERIVED from the `/rag` source files. When updating or adding content:

1. First check /rag directory for existing source material
2. Extract relevant information from /rag files
3. Restructure it into the format required for RAG (separate files per category)
4. Write in Ursa's voice (first-person, conversational)
5. Add proper YAML frontmatter metadata

This ensures consistency between the comprehensive source documentation and the RAG-optimized content structure.

Directory Structure

_content/
├── personal/          # Information about Vansh (personal context)
│   ├── bio.md
│   ├── skills.md
│   ├── experience.md
│   └── interests.md
└── projects/          # Project-specific information
    └── portfolio-website/
        ├── overview.md
        ├── tech-stack.md
        ├── challenges.md
        ├── outcomes.md
        └── links.md

Frontmatter Schema

All markdown files MUST include YAML frontmatter with the following structure:

---
type: personal | project
category: bio | skills | experience | interests | overview | tech-stack | challenges | outcomes | links
projectId: portfolio-website  # Required for type: project only
lastUpdated: YYYY-MM-DD
tags: [optional, tags, here]
source: /rag/personal/personal.md  # Reference to source file in /rag directory
---

Note: The source field documents which file in /rag the content was derived from, enabling traceability and making it easier to update content when source files change.

Writing Guidelines

CRITICAL: All content must be written in first-person ("I", "my", "me") AS Vansh, not ABOUT Vansh.

Follow the Ursa Personality Guide (docs/ursa-personality-guide.md):

• Tone: Conversational, authentic, passionate
• Voice: Strongly first-person
• Vocabulary: Clear, direct with informal touches
• Narrative flow: Stories, not bullet points
• Emojis: Strategic use (max 1-2 per document)

Content Requirements

Personal Content (300+ words per file):

• Bio: Personal story, what drives you
• Skills: Technical expertise with context and examples
• Experience: Work history as stories with impact
• Interests: Passions, side projects, learning journey

Project Content (200-600 words per file):

• Overview: What it is, why you built it, vision
• Tech Stack: Technologies used and why
• Challenges: Problems solved, solutions found
• Outcomes: Results, impact, lessons learned
• Links: Context for demos, GitHub, screenshots

RAG Integration

These files will be:

1. Processed by the data ingestion script (Epic 4 Story 4.2)
2. Converted to vector embeddings
3. Stored in Supabase vector database
4. Retrieved contextually during user queries

The type and projectId metadata enable context-aware filtering:

• type: personal → Used for hero section queries
• type: project + projectId → Used for project-specific queries