Lead Software Engineer / Architect — Agentic AI Platform

Type: Part-time based on specific project deliverables Location: Remote WHAT WE'RE BUILDING See www.juliet.space We're building Juliet, an AI that runs marketing end to end. Our users are marketers, founders and CEOs, growth leads, agencies, fractional execs, and SMBs, not developers. They tell Juliet the goal. She plans, writes production code, and ships real marketing: conversion-optimized websites, launch assets, campaigns, audits, autonomously. That's the engineering problem in one line: the humans in the loop can't read code, so the agent has to get it right on her own: plan, build, self-correct, recover, ship. Under the hood it's a browser-based studio backed by cloud sandboxes, a real-time streaming pipeline, and a LangGraph agent working across 83 tools and 63 skill modules. Small team, big ambitions. You'll ship things users touch daily, not write JIRA tickets about them. THE ROLE We're hiring one architect-level engineer to lead Juliet's technical execution. You'll own the architecture end to end, set direction for a small dev team, review what ships, and still write code every day. A player-coach, working directly with the founder. Your calls on agent behavior, system design, and infrastructure become the product. THE STACK Frontend: Next.js 15, React 19, TypeScript, Zustand, CodeMirror 6, XTerm.js API: NestJS 11, Supabase, Redis, PostgreSQL, Server-Sent Events AI agent: Python 3.11, LangGraph 1.x + LangChain, Anthropic / Google / OpenAI model providers Infra: Modal cloud sandboxes, Docker, Netlify deployments Monorepo: Turborepo, pnpm WHAT YOU'LL WORK ON Agentic AI workflows: design and improve multi-step LangGraph agent graphs that plan, write code, run commands, and self-correct across 83 tools, 63 skill modules, and a 30+ stage middleware pipeline Real-time streaming architecture: SSE pipelines that relay agent actions from the Python backend through NestJS to the browser, with reconnection, health monitoring, and interrupt handling (plan approvals, clarifying questions) Agent execution environments: sandbox lifecycle management on Modal: file sync, terminal I/O, command execution, and live preview with per-asset esbuild bundling State and memory systems: LangGraph Postgres checkpointers, middleware-injected context (goals, design docs, memory anchors), and conversation summarization Marketing intelligence and asset engines: public AEO, CRO, and brand-perception audit pipelines (multi-engine LLM probing, streamed reports) and a 20-format marketing asset studio Full-stack product features: collaboration (roles and permissions), the Netlify deployment pipeline, image handling, subscription/quota systems, and onboarding flows WHAT WE'RE LOOKING FOR Must-have: 8+ years of professional software engineering experience. This is a senior seat. Most people who thrive in it have closer to ten, including time as a tech lead or architect who owned a system end to end. Both worlds on your resume: engineering rigor inside a large company AND 0 to 1 ownership at an early-stage startup Strong TypeScript and Python. You're comfortable shipping production code in both. Production agentic experience. You've built and operated LLM agent systems with LangGraph, LangChain, or similar: agent graphs, tool use, state management, prompt engineering, evals. Well beyond calling a chat endpoint. AI-native development workflow. You drive Claude Code and Codex proficiently as everyday tools: agentic loops, skills, parallel sessions. You have opinions about working with coding agents because you do it daily. Experience with React/Next.js and building complex, stateful UIs Familiarity with real-time systems (SSE, WebSockets, streaming APIs) Architect-level system design. You can own how data flows across four services and defend every tradeoff. Strong plus: Background in developer tools, IDEs, or coding platforms Container runtimes and sandboxed execution (Modal, E2B, Firecracker) NestJS or similar Node.js backend framework experience Depth in PostgreSQL, Redis, and Supabase LLM observability and evals (LangSmith or similar) Exposure to marketing or growth products WHY THIS ROLE IS DIFFERENT You set the architecture. This isn't a feature factory, and you won't be implementing someone else's design doc. Technical execution is yours to lead. AI-native from day one. The agent isn't bolted on. She IS the product. You'll work at the intersection of LLM capabilities, agent infrastructure, and product design for non-technical users. Depth over breadth. The system spans cloud sandboxes, streaming infrastructure, agentic AI, and full-stack web. An IDE's internals where the only developer is the AI. There's always a hard problem to solve. Small team, big leverage. Your code ships fast and reaches users immediately. No layers of approval. HOW TO APPLY Send us: 1. Your resume 2. A short note on the most complex agentic system you've shipped: what broke, and what you'd redo 3. Your AI coding setup: which tools you drive daily (Claude Code, Codex, other) and how 4. A link to something you've built using agents or AI agentic system