Remote (US) · Full-time

AI Systems
Engineer

Build the AI that delivers our products. Agent workflows, automation pipelines, orchestration systems. Built once for Lynton, deployed again for clients.

About Lynton

Lynton spent 16 years as a HubSpot partner. 2,000+ companies, 50+ industries. We know SaaS from the inside: what it costs, how it locks companies in, where the architecture breaks down.

Two years ago, we saw what was coming. AI didn't just change the feature set. It changed the math. We left the partnership and started building what comes next. Today we help mid-market companies replace legacy SaaS with AI-native infrastructure they actually own.

The role

We sell AI-native website blueprints and sovereign stack blueprints as digital products, delivered primarily by AI. Someone has to build the AI that delivers them. That's this role.

The founder is building the first generation of autonomous AI agent teams, the delivery pipeline, and the automation frameworks right now. You inherit that foundation and make it production-grade, then extend it: new product types, deeper assessments, smarter automation, more reliable orchestration.

Build it once, ship it twice. The same AI systems you build for Lynton's internal operations get deployed for clients. The agent workflows that generate our blueprints become the agent workflows we implement for mid-market companies replacing their SaaS infrastructure.

What you'll do

  • Build and maintain the AI product delivery pipeline: the system that turns a client's website URL into a comprehensive technology blueprint
  • Create and orchestrate AI agent workflows for internal operations and client delivery
  • Extend existing AI-powered tools (website assessment engine, quote tool, blueprint generator)
  • Build the automation layer that makes Lynton's B2B ecommerce model work at scale with minimal human intervention
  • Design prompt architectures and agent coordination patterns that produce consistent, high-quality outputs
  • Integrate AI capabilities into sovereign stack deployments: intelligent CRM workflows, automated reporting, AI-assisted content operations
  • Monitor, evaluate, and improve AI system performance across output quality, cost, latency, and reliability

Who you are

  • 2–4 years building production systems with AI APIs (Anthropic Claude, OpenAI, or similar)
  • Strong Node.js or Python skills. You write real applications, not notebook demos
  • Prompt engineering beyond single-shot completions: multi-step chains, agent loops, tool use, structured outputs
  • You've built systems that run in production with real users, not just prototypes that work in a demo
  • Comfortable with API integrations, webhooks, and connecting multiple services into coherent workflows
  • You think about reliability, cost, and failure modes, not just "does it work once"
  • Comfortable with ambiguity. The field is moving fast and the playbook is being written in real-time

Bonus

  • +Vector databases, RAG patterns, or document processing pipelines
  • +Workflow orchestration tools or frameworks (LangChain, CrewAI, custom agent frameworks)
  • +Web scraping, data extraction, or content analysis at scale
  • +Familiarity with Astro, React, or modern web frameworks

Our AI stack

Primary AI

Anthropic Claude (Sonnet, Opus), structured outputs, tool use

Development

Cursor, Claude Code, custom agent workflows

Orchestration

Custom Node.js pipelines, Cheerio for web scraping, document processing

Infrastructure

Hetzner + Coolify (self-hosted Docker), Cloudflare CDN, PostHog

This stack is evolving. You'll have a voice in where it goes.

Why this role

  • Build the engine. You're not adding AI features to someone else's product. You build the AI systems that ARE the product, then deploy them for clients.
  • Ship, don't theorize. Not a research role. Every agent workflow, every automation pipeline, every prompt architecture goes into production.
  • Ground floor with credibility. 16 years of enterprise relationships, a legacy client base generating revenue, and a market thesis the numbers are proving right.
  • No gatekeepers. Small team. No ML platform team blocking your deployments. No six-month review cycles. You build it, it ships.
Express interest

We're building
the pipeline.

This role isn't our immediate hire, but we're actively building the pipeline for when it is. If you're a strong fit, we want to know you exist before we're racing to fill the seat.

Four questions designed to surface what resumes can't. Specific answers beat polished ones. Generic AI-generated responses won't make it through.

Step 1 of 3

Tell us who you are

The basics. We'll only use this to follow up on your application.

City, state. Remote is fine, just helps us know your timezone

Step 2 of 3

Show us your work

Provide at least one: LinkedIn URL, resume PDF, or both. A portfolio link is optional but helps us see your design taste.

Required if no resume uploaded

Optional if you provided LinkedIn. Max 10MB.

Optional but recommended. Show us what you've built

Step 3 of 3

Tell us how you work

Be specific. Generic answers don't help either of us. A few sentences each is plenty. We're looking for substance, not length.

Architecture, tools, how it handles failures, how many users it serves.

Multi-step chains, tool use, structured outputs, evaluation.

Give us a real example of optimizing or debugging an AI system.

Link to it if possible.

Applications are screened by AI against the role criteria. Real humans review the recommendations.