Xavion Mirchandani

When I was 13, I was fascinated by AI but frustrated by how fragmented it was. Every model lived behind its own signup, its own interface, its own subscription — and there was no easy way to compare them. I found myself juggling browser tabs, pasting the same prompt into ChatGPT, Claude, Gemini, and Perplexity to figure out which gave the best answer.

So I built the tool I wanted. Over one summer — more than a thousand hours of it — I taught myself React, Next.js, APIs, authentication, and deployment while building Xavion AI, a platform that puts 30+ models behind one interface. Partway through, I realized the product had grown into a patchwork of separate pages, so I scrapped the interface and rebuilt it as one seamless experience. Undoing weeks of my own work was painful, but it taught me that design decisions compound just as fast as code does.

In the summer of 2026 I joined Preface, a global technology-education company, as a software engineering intern in Singapore. As the only engineer on the local team, I owned real systems end-to-end: a three-agent Python pipeline that automates client outreach in production — drafting personalized emails, reading replies, and keeping my boss's calendar reconciled — and Preface 2559, an internal CRM where AI agents keep the pipeline current but every outward-facing action waits in a queue for human approval. Working on systems that touch real clients taught me more about trust, guardrails, and data integrity than any personal project could.

Away from product work, I compete — 1st place as team captain at Lockheed Martin Code Quest (Asia), and I'm currently competing in the USA AI Olympiad. I built a 360TB TrueNAS home server from parts, and I've managed my own investment portfolio since I was ten, tracking every trade and its reasoning.

I grew up in Singapore and now board at Blair Academy in New Jersey, where I'm on the AI Student Advisory Council helping shape how the school adopts AI.