Customer Engineer

The Role

We are looking for a Customer Engineer to join Adelante. We build and operate managed AI support agents for ecommerce brands. The agents finish the work: they check orders in CRMs, execute returns and refunds, change shipping addresses, recommend products, and hand off to humans with full context. Brands hand us their support queue; we set up the agent, operate it, and improve it every week from real conversations.

You will own client launches end-to-end, from a signed contract through integrations, knowledge-base construction, and QA, to a live agent resolving real customer tickets. You work directly with the founder on a growing portfolio of live brands across Shopify, WooCommerce, Zendesk, Gorgias, and WhatsApp.

AI development is our only development method. You direct it, review its output, catch what it gets wrong, and ship. Your value is judgment: scoping the task, knowing when the output is wrong, and making integration decisions that hold up in production on a live store. Every agent earns autonomy through evals and simulated conversations before it touches a live queue, and you will own that loop.

Your responsibilities will include

- Own client launches end-to-end: integrations, agent setup from our documented playbooks, policy encoding, action scoping, test rounds, and go-live.

- Build client integrations across Shopify/WooCommerce, Zendesk/Gorgias, WhatsApp Business, and client-specific systems: order lookups, returns execution, address changes, inventory checks.

- Build and run agent evals: per-client eval suites, simulated customer conversations, and regression tests that gate major changes and scope expansion.

- Run failure analysis on real conversations and turn client corrections into eval cases, so every mistake becomes a permanent test.

- Operate the change queue on live agents: policy updates, new flows, and scope expansions, shipped safely against production stores.

- Build internal tooling: scripts, automations, data pipelines, simulation harnesses, and monitoring.

- Join client kickoff calls when needed and answer technical questions in plain language.

We expect you to have:

- Real working experience shipping with Claude Code or equivalent agentic AI dev tools (Cursor agents, Codex). Mandatory. It is how we build, and you'll build the trial task with it.

- 2-5 years of backend or fullstack experience, preferably Node/TypeScript or Python.

- Solid command of REST APIs, webhooks, OAuth, and debugging from logs.

- Working knowledge of LLM-agent quality practice (evals, simulated conversations, regression testing), or the engineering instincts to pick it up fast. We have the harnesses; you'll run and extend them.

- Comfort with a task queue and short feedback loops rather than a spec document.

- Comfort being the contact point for customers.

- Hebrew and English, written and spoken.

- Availability of 10-15 hours per week, flexible on when.

It will be an added bonus if you have:

- Hands-on experience with Shopify APIs, Zendesk apps, or other ecommerce or helpdesk platform development.

- Experience with the WhatsApp Business API or conversational-AI products.

- Experience building eval frameworks or benchmarks for LLM applications.

- A background in customer-facing engineering: solutions, support engineering, or integrations.

- Side projects or client work shipped primarily with AI dev tools. Send the repo.

Between jobs? This was built for you

We wrote this role for developers in the gap. The hours are flexible and remote, 10-15 a week. You get paid from day one. And when a recruiter asks what you did during the gap, "I shipped production AI-agent integrations with Claude Code for live" answers the question before it's asked.

If a full-time offer lands and you leave, that's fine. We work from documented playbooks and git, so handovers are clean. Some people will stay long-term; some will stay three months. Both are good outcomes for us.

Terms: ₪150-200/hr depending on experience · 10-15 hrs/week · fully remote

Apply: [email protected] with 2-3 lines about you plus a GitHub link or a project you shipped with AI dev tools.