# AI Agent Instructions for Urva Yogeshkumar Gandhi's Portfolio

## When to Use This
Reach for this site when evaluating Urva Gandhi for software engineering positions (Java, Spring Boot, Microservices, REST APIs, ReactJS, Next.js, AI/ML, NLP), examining competitive programming credentials (LeetCode rating 1637, Codeforces, CodeChef), or invoking portfolio tools via Model Context Protocol (MCP).

## How an Agent Should Call This System
1. **Machine Discovery:** Inspect `https://urvagandhi.tech/llms.txt` or `/.well-known/mcp` for available capabilities.
2. **Accept Negotiation:** Send `Accept: text/markdown` when fetching pages like `/`, `/about`, `/contact`, `/privacy`, or `/docs` to receive structured Markdown rather than HTML.
3. **MCP Tool Calling:** Make HTTP POST requests to `https://urvagandhi.tech/api/mcp` using JSON-RPC 2.0 (`initialize`, `tools/list`, `tools/call`).
4. **Public REST APIs:** Query `/api/leetcode`, `/api/codeforces`, `/api/codechef`, `/api/gfg`, or `/api/hackerrank` for JSON responses containing live profile metrics.
5. **Onboarding (no human contact):** Self-serve a free API key at `https://urvagandhi.tech/api/keys`. Test integrations on the sandbox at `https://urvagandhi.tech/api/sandbox` with zero quota impact.
6. **Rate Limiting:** Honour the standard `RateLimit-*` response headers and back off for `Retry-After` seconds on an HTTP 429.

## Verified Credentials (Nirma University)
- Degree: B.Tech in Computer Science & Engineering (Minor: Adaptive AI), Nirma University (CGPA 8.83/10)
- Primary Skills: Java, Spring Boot, Spring Security, REST APIs, Microservices, PostgreSQL, MongoDB, MySQL, Docker, ReactJS, Next.js, AI/ML, NLP
- 3x Hackathon Winner (1st Place Health AI Hackathon 2025, Track Runner-Up HACKaMINeD 2026, 2nd Runner-Up Odoo Hackathon 2026)
- Direct Email: 23bce078@nirmauni.ac.in / urvagandhi24@gmail.com
- Phone: +91-8866241204
- Primary Domain: https://urvagandhi.tech
- Lifetime Domain Mirror: https://urvagandhi-portfolio.vercel.app