What EmailPilot is
EmailPilot.ai is a retention-intelligence SaaS for e-commerce brands. It connects to a brand's Klaviyo and Shopify accounts, ingests customer and campaign data, and produces campaign calendars, briefs, and strategy recommendations that a retention team can act on. It's built on Google Cloud (Cloud Run, Firestore, Cloud SQL, Cloud Build, Secret Manager), runs Claude models under the hood, ships an MCP server, and is in the middle of a Klaviyo Marketplace submission.
It was built by the founder — an 18-year email marketing operator who runs a Klaviyo Master Platinum agency — working with AI coding agents. It works. It also has the shape you'd expect from a founder-built, single-developer codebase: multiple services, some duplicated paths, hard-won operational invariants that live in docs and cloudbuild files rather than in tests.
You would be the second person ever to work in this codebase.
Why this role exists
The company is moving from "founder builds it" to "founder sells it." For the next two years the founder's job is customers, and yours is everything that keeps the product running, integrated, and shippable. You are not being hired to build the roadmap. A feature freeze is in effect until the first cohort of paying customers has told us what actually matters — the first 90 days are about reliability, the marketplace listing, onboarding, and fixing what customers hit.
If you want a greenfield build, this isn't it. If you like taking a working system someone else built, understanding it deeply, making it boring and dependable, and then earning the right to shape what comes next — keep reading.
What you'll own
- Reliability and infrastructure. Cloud Run services, Firestore, Cloud SQL, Cloud Build pipelines, monitoring and alerting, cost. Some services have deliberate single-instance and warm-instance constraints with a history behind them; you'll learn why before you change them.
- Klaviyo Marketplace submission. OAuth/PKCE flow, token storage, review requirements, and whatever Klaviyo comes back with.
- Onboarding flow. A new customer connects Klaviyo and Shopify and gets a usable campaign calendar in under 30 minutes. It mostly works. Make it work every time.
- Integrations. Klaviyo, Shopify, Asana, Slack, the MCP server, and the LLM layer (prompting, output validation, cost control).
- Codebase hygiene. Consolidate without a big-bang rewrite. Tests where they pay for themselves. Documentation that a third engineer could onboard from.
- Customer-driven fixes. When a paying customer hits something, you're the one who fixes it — usually the same day.
The founder keeps product decisions, pricing, and every customer conversation. You'll be in those conversations when a technical answer is needed.
How we work
- AI-first, engineer-owned. Claude Code and similar agents are the default way work gets done here, and you're expected to use them heavily. You're also expected to read every line they produce, know when they're wrong, and never ship something you can't explain. If your instinct is "the model wrote it, so it's fine," this is the wrong role. If your instinct is "the model wrote it, so I'll review it like a junior's PR," this is the right one.
- Async by default. Slack for conversation, Asana for work, written proposals for anything that changes architecture. The founder is technical enough to review your work and busy enough that he shouldn't have to.
- Correct over fast. When unsure, stop and ask. A slower right answer beats a fast wrong one, especially in production.
- Small team, real stakes. There is no QA team, no on-call rotation, no one else to catch it. That's the trade for real ownership.
What we need
- 5+ years shipping and operating production web services, at least 2 of them where you were the primary owner of a system in production.
- Google Cloud in production: Cloud Run, Firestore, Cloud Build, Secret Manager, IAM. Cloud SQL or Postgres.
- You've inherited at least one sprawling codebase you didn't write and made it better without rewriting it. Be ready to talk about what you deleted.
- Daily, fluent use of AI coding tools with demonstrated judgment about their output.
- OAuth and third-party API integrations — building them, debugging them, handling token lifecycle.
- Clear written English. Most of your communication with the founder will be written and asynchronous.
- Reliable availability with at least 4 hours of overlap with US Central time.
Nice to have
- Klaviyo or Shopify API experience, or email/SMS marketing domain knowledge.
- Building LLM-powered products: prompt design, structured outputs, evals, cost management.
- MCP (Model Context Protocol) servers.
- BigQuery, data pipelines, or a RAG system you've kept running in production.
Engagement
- Independent contractor, monthly retainer, paid via Wise. Any country.
- Starts with a paid 30-day trial on a scoped set of real tasks. If it works for both sides, it converts to an ongoing retainer.
- Compensation: $25-$45 USD /hr. State your rate and expected hours in your application.