Full Stack Engineer

About Kaitoi

Kaitoi is an AI-native orchestration platform for creative professionals. It is agentic across the entire surface (creative collaboration, execution, guidance, and live extension of the software itself), not just a chat panel bolted onto a fixed feature set. A single project flows between a conversational moodboard, a node graph, a workflow view, a storyboard and a timeline, all interconnected and live.

It carries a production from first idea to rough cut, exports cleanly to professional editors, and plugs into existing pipelines through API, MCP, a built-in local inference service, and an automatic recipe system that turns HuggingFace or GitHub model repos into portable, repeatable workflows.

We are building one of the most ambitious creative computing platforms out there with a very small team. The work spans real-time execution systems, sandboxed Python runtimes, graph-based creative tooling, collaborative interfaces, and the infrastructure that makes all of it production-grade.

The role

You will own meaningful slices of the product across the stack. Not just shipping features, but shaping how they work. One week you might be tightening WebSocket session routing in the queue execution path; the next, redesigning a node inspector in the editor; the next, hardening a new authentication flow.

We expect engineers who think about performance, reliability and security as part of the work, not as a follow-up ticket.

If you prefer one side of the stack, that is fine too. Just say so. We are open to specialists who want to go deep on either the frontend or the Python backend.

We care a lot about thoughtful engineering and sustainable architecture, even while moving quickly. The scope is ambitious, but we try hard to keep the system understandable.

What you will do

• Build features end-to-end across the TypeScript frontend and Python backend
• Design and ship surfaces of the node graph editor, timeline view, asset library, and assistant experience
• Extend the execution engine, per-user sandboxing, and worker/queue pipeline
• Improve the data model and storage paths (document database, object storage, signed URLs, multipart upload)
• Strengthen authentication, authorization, and tenant isolation
• Profile and optimize hot paths: graph execution, large-asset I/O, real-time previews
• Help scope, plan, and review work alongside the rest of the team

What you bring

• Strong production experience across at least one modern frontend framework and one backend language; comfort working in TypeScript and Python is ideal
• Deep familiarity with several of: a modern frontend framework (Svelte, React, Vue, Solid, SolidStart, Next.js, Nuxt), a Python web framework (FastAPI, Django, Flask, Starlette) or Node.js (Express, NestJS, Hono), document or relational databases (MongoDB, Postgres, DynamoDB, SQLite), WebSockets and real-time protocols, object storage and CDN delivery, queue-based or worker architectures, container orchestration
• A real bar for performance, security and reliability. You reach for profilers, you write threat-aware code, and you reason about failure modes before they bite
• Comfort working in a fast-moving codebase: reading existing patterns first, making focused changes, and avoiding speculative abstraction
• Strong written communication and the ability to operate independently in a remote, async setup

Nice to have

• Experience with sandboxed code execution (subprocess isolation, virtualenvs, resource limits)
• Visual editor / canvas / graph UI experience (SVG, custom layout, large-tree rendering)
• Hands-on experience with a major cloud provider (AWS, GCP, or Azure): container orchestration, managed queues, object storage, CDNs
• Experience integrating LLMs / agent frameworks / MCP

How we work

• Fully remote, global. Async-first written communication.
• Small team, broad ownership, short feedback loops.
• We ship to production frequently and value working software over large process.

How to apply

Send a short note about what you have built and a link to code or product we can look at. Pointers to one or two things you are proud of go further than a full résumé.