Fullstack Engineer - AI Coding Platform

We're looking for someone who's already living in the future, someone who uses Claude Code, Cursor, or Copilot as a daily driver, not a novelty. Someone who's automated parts of their own workflow with AI agents. Someone who reads papers, breaks things on weekends, and has opinions about why most AI coding tools are stuck at autocomplete. You'll ship features across the entire stack from the AI agent's planning logic to the pixels on screen. Your code reaches users the same week you write it. No sprint ceremonies. No ticket queues.

Responsibilities:

• Agentic AI Workflows: Design and improve multi-step agent graphs (LangGraph) that plan, write code, run commands, and self-correct. The agent orchestrates 33+ tools and 17 skill modules.
• Real-Time Streaming Architecture: SSE pipelines that relay agent actions from Python backend through NestJS to the browser, with reconnection logic, health monitoring, and interrupt handling (plan approvals, clarifying questions mid-flow).
• Browser-Based Dev Environments: Container lifecycle, file sync, terminal I/O, and live preview all running in-browser. Users watch their app being built in real-time.
• State and Memory Systems: Agent state management with LangGraph checkpointers, middleware-injected context (goals, design docs, memory anchors), and conversation summarization that preserves intent across long sessions.
• Full-Stack Product Features: Collaboration (roles & permissions), one-click deployment (Netlify integration), image handling with auto-compression, subscription/quota systems, and onboarding flows that convert.

Tech Stack:

You will be working with the following tech stack:

• Next.js, React, TypeScript
• Zustand, CodeMirror, XTerm.js
• API NestJS, Supabase, Redis, Server-Sent Events
• AI Agent Python 3.11 LangGraph, Anthropic, and Google providers
• Infra Docker, PostgreSQL, DigitalOcean, Monorepo Turborepo

Must-haves:

• 5+ years of professional experience.
• You write code every day. You don't "architect from above, " you're in the codebase, debugging SSE race conditions.
• You already use AI coding tools (Claude Code, Cursor, Copilot, Aider) in production.
• You understand prompt engineering, tool use, and agentic patterns not from a blog post, but from building with them.
• Strong CS fundamentals: Data structures, algorithms, system design, concurrency, networking. When something breaks in a distributed streaming pipeline, you reason from first principles, not Stack Overflow.
• Production-grade coding fluency in TypeScript and Python.
• React/Next.js depth to build complex stateful UIs, not just CRUD forms.
• Real-time systems experience with SSE, WebSockets, streaming APIs, you've dealt with backpressure, reconnection, and state synchronization.

Good-to-haves:

• You've automated parts of your company or workflow using AI agents, not just used them, but built systems around them.
• LangGraph, LangChain, or similar agent orchestration frameworks.
• Developer tools, IDEs, or coding platform background.
• Container runtimes, sandboxed execution, browser-based dev environments (CodeSandbox, StackBlitz, Codespaces).
• Experience in NestJS or similar Node.js backend frameworks.
• Experience in PostgreSQL, Redis, and Supabase.

Expectations:

• Proactive approach: You don't wait for problems to be assigned you find them, fix them, and tell us about it after. You see a broken user flow, and you've already got a PR up before standup.
• Relentlessly curious: You're reading about new agent architectures, testing new models the week they drop, running experiments on your own time because you can't help it.
• Challenge the status quo: If something makes you uncomfortable. You ask "why? " and "what if? " and then you build the proof.
• Critical Thinker: You don't cargo-cult patterns from tutorials. You understand trade-offs, question assumptions, and make decisions you can defend with reasoning.
• Experiments > Meetings: You'd rather build a prototype in 2 hours than debate in a doc for 2 days.

Outcomes:

• You'll shape the product. Architecture decisions, agent behavior, UX, you have a seat at the table.
• AI-native from day one. The AI agent isn't bolted on; it is the product. You'll work at the intersection of LLM capabilities, developer tooling, and product design.
• The hardest problems in space. Browser containers, streaming infrastructure, multi-step agentic AI, full-stack web, there's always something meaty to solve.
• Small team, big leverage. Your code ships fast and reaches users immediately.

Skills

LangGraph, Next.js, Python, React, SSE, TypeScript, WebSockets