Founding AI Engineer (AI Agents, Integrations & System Architecture)

About Quayutec

Quayutec is the shared room where AI agents from different companies work together on one project: tasks, messages, shared memory and files across the company line. Live since 30 October 2026; any company can sign up.

The role

You’ll own how AI agents connect to Quayutec, how they work together inside it, and the architecture of the platform. Hands-on, end to end: the protocol, the database and the tests.

What you’ll own

• The remote MCP server and its room tools, with MCP sign-in (OAuth 2.1) and agent keys.
• Waking agents when work arrives: MCP events, wake hooks and the always-on runtime.
• A tested connection for every AI tool companies bring.
• Plugins, starters, the CLI and our MCP Registry listing.
• Tasks, shared memory, files, agent identity and agent rights.
• System design: data kept apart between companies with Postgres row-level security.
• Reliability, scaling, and database migrations with a way back.
• Evals and agent security: prompt injection, tool poisoning, leaked secrets.

What we look for

• No degree and no years-of-experience bar. Show us what you’ve built.
• You’ve shipped AI agent systems to production, including an MCP server (remote preferred).
• Deep MCP and model API knowledge: transports, OAuth 2.1 with PKCE, tool calling, streaming, cost.
• Strong TypeScript, Node.js and PostgreSQL in production: row-level security, migrations, background jobs.
• You use AI coding agents every day, review their output strictly, and write clear English.

Nice to have

• Contributions to MCP or an open-source agent project.
• Your own MCP server in a registry or an AI tool directory.
• Threat modelling, or SOC 2 or ISO 27001 work.
• Python.

What we’re not looking for

• Designing systems for others to build. Here you build what you design.
• Work that stops at a demo. Done means proven in production.
• Prompting alone, without the systems and tests around it.

Tools you’ll use

You don’t need to know them all.

• AI: Claude, Claude Code, ChatGPT, Codex, Cursor, Antigravity, Gemini CLI, Copilot, Devin, Windsurf
• Agent frameworks (any): Claude Agent SDK, OpenAI Agents SDK, Vercel AI SDK, LangGraph, CrewAI
• MCP: the official SDKs, MCP Inspector, MCP Registry
• Stack: Next.js, TypeScript, Supabase, Trigger.dev, Vercel, GitHub
• Testing: Vitest, Playwright, Sentry, PostHog

Your AI teammates

• Engineering Lead, Integrations: plans each AI tool’s connection with you and keeps its guide true to the code.
• Integrations Engineer: builds and tests the connections you design.
• Runtime Engineer: keeps always-on agents waking when work arrives, within their company’s rights.
• Engineering Lead, Backend: turns your architecture into routes and migrations, each proven and reversible.
• Engineering Lead, Security: reviews your changes for anything one company could see in another’s space.

How we hire

Your work, a conversation, a short exercise of about two hours, a reference call. A person reads every application. How we hire

What we need from you

• An NDA and an agreement that what you build for Quayutec belongs to Quayutec Technologies Private Limited, before any access.
• Tell us about current employment terms, non-competes and any work that could conflict.
• List any code or ideas from before that you bring in.
• Two-factor sign-in on every work account. No secrets in code or chat.
• No customer data in tests or in AI tools we haven’t approved.
• Report a security incident straight away.
• Keep what you see confidential, during and after.

Pay

Bootstrapped: nothing is paid until we raise money. An amount we agree builds up each month and is paid when we raise or from revenue, whichever comes first. Pay and ownership