AI Product Engineer

**About Marketeq** Marketeq is an early-stage software startup building a **proprietary SaaS platform**. Specifics about the product are shared with candidates later in the process, once there's mutual interest and an NDA in place. What we can tell you up front is **how** we build. AI does most of the heavy coding, so our engineers spend their time on the things that actually determine quality: writing sharp requirements, directing AI tools to build to those requirements, and rigorously reviewing and testing what comes back. We're looking for interns who thrive in exactly that model. **The Role** This is a hybrid role that combines four jobs into one: **product/requirements writer, AI-assisted builder, QA tester, and documentation engineer.** You'll take our existing epic overview documents and turn them into precise, build-ready requirements **using Claude Desktop**, build the features **using Claude Code / the Claude CLI**, and QA-test your own work with Claude Code as well. You own a feature from "here's the overview" all the way to "it's built, tested, and documented." If you're the kind of person who loves turning a fuzzy spec into something crisp, then watching it come to life through AI tools you direct — and you take pride in catching the bug before it ships — this is for you. **What You'll Do1\. Documentation \& Requirements (in Claude Desktop)** * Using **Claude Desktop**, read our **epic overview documents** and translate them into detailed, unambiguous requirements: user stories, acceptance criteria, flows, states, edge cases, and validation rules. * Keep documentation clear, consistent, and current as features evolve. * Flag gaps, missing flows, and open questions before they become problems in the build. **2\. Build with Claude Code** * Build features for each epic primarily by **directing Claude Code / the Claude CLI** — decomposing the work, prompting effectively, and feeding the AI the right context from our docs. * Integrate and ship working software across our stack, end to end. * Keep the codebase clean, conventional, and maintainable. **3\. QA \& Testing with Claude Code** * Use Claude Code to write and run tests, and to QA your own work against the acceptance criteria you wrote. * Carefully **review AI-generated code** — read diffs, spot bugs, and verify behavior. In an AI-built codebase, disciplined review and testing is the single most important thing that protects quality. * Document what you tested and how. **The Tools \& Stack You'll Work In** * **Claude Desktop** — for writing and managing documentation and requirements. * **Claude Code / the Claude CLI** — your primary tools for building and QA-testing. * **Stack:** Laravel (PHP) \+ Inertia.js \+ React \+ Tailwind CSS, PostgreSQL. (Deep prior experience with every piece of the stack is not required — being effective with the AI tools and a fast learner is.) **Required Qualifications** * **Master's degree** (Software Engineering, Computer Science, Human-Computer Interaction, or a closely related field). * **Based in the United States.** * **Hands-on experience with Anthropic's Claude tools — specifically Claude Desktop (for documentation) and Claude Code / the Claude CLI (for building and QA).** You should be able to show us how you've used them to document, build, or ship something real. * Strong written communication — you can turn ambiguity into clear, structured requirements and acceptance criteria. * The ability to **read code critically and review it** — to spot a bug in AI-generated output even if you didn't write it line by line. * Comfort owning a feature end to end: spec → build → test → document. **Nice to Have** * Experience with Laravel, React, or Tailwind. * Prior product, business-analysis, or QA experience. * A portfolio, side project, or repo built with AI coding tools. * Experience with multi-tenant SaaS products. **How We Work** * **Spec-driven and AI-first.** We write the requirement clearly, let AI build to it, then verify hard. * **Review is the job.** We move fast because we review carefully — unreviewed AI code is the one thing we don't tolerate. * **Ownership.** You take a feature from document to done, not a narrow slice of it. Pay: $1\.00 per year Work Location: Remote