Ops Lab — AI demo by Tusharika Singh

Welcome to the Lexavra Infinology DevOps take-home demo!

tusharika@aws-mumbai:~/lexavra-demo$

Hi, I'm Tusharika. This page is the visible tip of the full stack from the assignment: a Next.js frontend on Vercel, calling an n8n workflow engine on AWS over TLS, answered by a self-hosted Llama 3.2 model served by Ollama — no third-party AI APIs. Every design decision, config and diagram lives in the GitHub repository.

Demo notes:

  1. The backend sleeps overnightIt runs on a personal AWS account and is live daily from 10:00 to 22:00 IST — deliberate cost discipline for a demo, not an outage.
  2. Small model, honest answers about itReplies come from llama3.2:1b on CPU. Give it a few seconds, and review anything important — small models make mistakes.
  3. Single-turn by designEach question is answered independently; the conversation history lives only in your browser tab.

Built by Tusharika Singh · view source & architecture on GitHub

Ask anything — for example “Explain what a webhook is in one sentence.”

Press Enter to send · Shift+Enter for a new line