AI Automation Engineer - Reformly · Remote · Full-Time

Compensation: $20k – $40k • No equity

Summary Location: Remote Type: Full-Time (this is an in-house build role, not a project-based engagement) Compensation: Base salary, based on experience — to be discussed

About Reformly Reformly helps women build confidence through Pilates equipment and 20-minute guided home workouts, combining physical boards with our own subscription app. Since launching in May 2025, we've generated $4.56M in revenue and sold 30,000+ boards. In February 2026 we launched our custom subscription app, now at 3,600 active subscribers and over $40K MRR.

We're currently wanting to invest in building our own AI-automation systems in-house instead of outsourcing work AI can now do better, cheaper, and faster.

The Role

We're hiring an AI Automation Engineer to take full ownership of building the production systems that replace three current bottlenecks — static ad creative, customer support, and content/UGC production — while also strengthening our core Shopify and app stack.

This isn't a "call an LLM API and ship a demo" role. We want someone who builds systems that keep working when third-party APIs fail, models return unexpected output, jobs take minutes to complete, and the same webhook arrives twice — and who operates with full autonomy and ownership, thinking proactively about what to automate next without being asked.

What you'll own

• Full ownership of our AI-automation build-out, starting with three concrete systems:
  • An internal static-ad-generation system: competitor ads + our brand assets + base prompts → on-brand ad creative, revisable via prompting, eventually closing the loop with our Meta Ads performance data
  • An in-house AI customer support layer (LLM + tool-calling) that resolves a large share of tickets and can take real actions, replacing our current Gorgias-based setup
  • An AI-assisted content/UGC generation system to reduce our dependency on sourcing, shipping boards to, and paying human creators
• Backend systems: APIs, webhooks, background jobs, and workflow orchestration that are resumable and observable, not fragile scripts
• Production AI engineering: structured output/tool-calling, RAG (pgvector), confidence thresholds for when AI should act vs. escalate to a human, and cost/token accounting per pipeline
• Ongoing: subscription/MRR/churn dashboards, shop and marketing analytics, video hosting cost optimization, video-editing tooling, and general full-stack flexibility across our Shopify store and Mobile app
• Spotting and pitching new automation opportunities — without being asked Fast response to breakages or failures, including during high-stakes periods

What We're Looking For

Must-haves:

1. Strong TypeScript/Node.js, ideally with NestJS or another DI-based framework
2. PostgreSQL (schema design, indexing, migrations, transactions)
3. Experience with real-world third-party APIs that are slow, rate-limited, paginated, or unreliable — plus webhooks done right (signature verification, idempotency, retries, dead-lettering)
4. Background jobs/queues (BullMQ, Redis, or equivalent)
5. Verifiable experience shipping LLM-powered features (OpenAI and/or Anthropic APIs) that real users interact with — not notebooks or demos
6. Structured output / tool-calling / JSON schema validation, including handling malformed or refused model responses
7. Solid React/TypeScript and Next.js (App Router); Tailwind
8. Proven autonomy — you've owned a system or codebase end-to-end without hand-holding
9. Proactive communicator — you share updates, risks, and ideas before we have to ask

Nice-to-haves:

1. RAG experience (chunking, embeddings, vector search — we use pgvector)
2. Workflow orchestration platforms (Temporal, Inngest, Trigger.dev) or hand-rolled resumable state
3. Media generation APIs (Higgsfield, Replicate, fal) — submitting jobs, polling, storing assets, handling retries
4. AI evaluation pipelines (golden test sets, regression testing)
5. Shopify, Stripe, Klaviyo or similar commerce/marketing APIs
6. React Router 7, Sentry, GitHub Actions, Docker-based deployment

How to Apply

We review applications primarily through a short video application — this lets us get to know you quickly and keeps the process moving fast for both sides.

Please submit:

• A 3–5 minute video (recorded via Loom preferred — phone or laptop recording is fine too, no need for production value) answering:
  • Walk us through one system you've built end-to-end that you're genuinely proud of — what was hard about it, and what decisions were yours (not a library default)
  • Tell us about a time you were given more ownership or responsibility than your role technically called for, or proactively spotted a problem before anyone asked you to look into it. What did you do?
  • Pick one of the three systems described above (ad creative, support, or content). In a couple minutes, tell us how you'd start thinking about it — we're not looking for a finished plan, we're looking for how you reason about an ambiguous, real problem
  • Describe a real situation where a third-party API or webhook gave you unreliable or unexpected behavior in production, and exactly what you did about it
  • What does your current availability look like (other commitments, capacity), and how do you handle being "on" when something breaks outside normal hours?
• A short written intro including:
  • Links to your work: GitHub, live products, portfolio, or case studies with real outcomes
  • Your current setup: full-time available, freelance, or currently employed elsewhere — and how much capacity you genuinely have for this
  • Your proposed compensation expectations
  • (Optional but appreciated): a short outline of how you'd approach your first 30 days in this role

Submission: Reply to this post / application form with your video link (Loom preferred — Google Drive or YouTube unlisted also fine) and written intro attached.

We review every application against a structured scorecard. Strong applicants will be invited to a live call.

Note: this is a high-ownership role for someone who wants to genuinely grow with the brand — not a hands-off, ticket-execution engagement.