IgniteTech Recruitment

The way most companies build enterprise software ages in a predictable way: large teams, long roadmaps, quarters to ship what should take days. When AI capability jumps, the org can't absorb it — because it was never built around AI. We rebuilt around it: AI is in our DNA, not bolted on. And we're a large enterprise organization with big customers that moves like a startup — cycles in weeks, not quarters. Our customers are Fortune 100 brands, so when our software breaks, it's their reputation on the line. The bar is enterprise-grade and fast.

This is a senior, full-stack engineering role on an enterprise B2B product. The skill that matters most here isn't typing code — it's building the harness: the tools, skills, context, and closed loops that let AI agents do whole units of work and verify their own output. In the AI era the harness is the product — the durable thing you build is the system that produces features, not any single feature — and you'll spend real time orchestrating multiple agents in parallel, each wired with the context and verification it needs. You own features end-to-end across every layer, you mentor more junior engineers, and you share the harness you build back so the whole team gets sharper, not just you.

What You Will Be Doing

Build the harness — the tools, skills, context, rules, and closed loops that let agents do whole units of work and verify their own output — so each class of work needs less of you than the last. This is the core of the role.
Orchestrate multiple agents in parallel across every layer (UI, API/GraphQL, backend, containerized services), each wired to the context and verification it needs, and ship enterprise-grade features end-to-end.
Multiply the team: share the harness, rules, and skills you build back so the gain compounds — and mentor more junior engineers, leading by example.
Own quality: the complete spec, no regressions, a clean PR review first-pass, good product taste, enterprise-grade.
Get out of the AI's way — your value is the system you build that makes agents deliver, not the lines you type.

What You Will NOT Be Doing

Writing code by hand, or sitting in 1:1 loops with an AI assistant. AI does the work; you build the loops that let it, and you own the result.
Relying on a handoff chain. You own your feature end-to-end — you don't punt the UI to a designer or lean on QA to catch your bugs. (You won't be expected to own whole-product infrastructure solo — you run your features end-to-end and lean on the team for the rest.)
Settling on one tool. You operate at the frontier and keep testing what's next.
Waiting for work or being managed. You're proactive and own outcomes.
Treating "it runs" or "merged" as "done." Done is the complete spec, tested, clean review, in a release a Fortune 100 customer can rely on.

Requirements

AI-DNA — it's how you operate, not a tool you reach for. You don't write code by hand or sit in 1:1 loops with an assistant. Your default is to get out of the AI's way: build the context, closed loops, and guardrails that let agents own work end-to-end, then direct and verify.
Multiplies through AI (1:many), contributes to the team (1:many²). You delegate real units of work to agents running closed-loop and in parallel — not turn-by-turn back-and-forth — and you share the leverage you build back so the team compounds.
Senior full-stack — Java + ReactJS. 4+ years of full-stack engineering with genuine production depth in both Java (backend) and ReactJS (front end); you own features across every layer — UI, API/GraphQL, backend, and containerized services.
Runs features end-to-end — you build, test, ship, and operate your own features without a handoff chain.
Heavy, advanced AI-coding & agent user — and a relentless experimenter. Power-user command of today's frontier agentic coding tools; never locked to one; constantly testing alternatives. Going deep on only one tool, or dabbling shallowly across many, are both red flags.
Sharp, current model judgment — which frontier model to reach for and why, what each is better at, what changed across recent releases.
Closed-loop quality discipline — you give AI and yourself ways to verify (tests, build/type checks, running the containerized app, browser checks for the UI) so speed never means regressions. No open-loop AI output.
Product taste — you think about the user and the experience, handle edge cases, and ship polished work without needing every detail spec'd.
Ownership & proactiveness — mission-driven, obsessive about quality, no babysitting; surface risks early and overcome blockers.
Enterprise-grade software experience — defensive coding, edge cases, validation, security, testing, and code review as defaults.
Working CI/CD knowledge — comfortable enough with CI/CD (e.g., GitHub Actions or similar) to troubleshoot a build failure; AWS familiarity a plus.
Mentors and reviews — you've guided or mentored other engineers and reviewed their code, even without a formal lead title.

Nice to Have

Experience building AI/LLM-powered product features (APIs, prompt engineering, agents, automation).
Prior work on enterprise community, social, or B2B SaaS products.
GraphQL APIs and containerized services at scale.
Building or contributing reusable internal tooling — rules, skills, MCPs — that measurably raised a team's output. AWS experience.

What You Will Learn

You'll do frontier AI-native engineering on a real enterprise product — learning how to get far more out of agents than most engineers think possible: closed loops, context-wiring, parallel autonomy, and the leverage that compounds. The AI-native way of working here is already in production.

Working Conditions

Fully remote, async-first, global. No office. Compensation: $100k/yr. We default to async — status updates over status meetings, a shared written source of truth, deep work on your own schedule — paired with a fast feedback loop where product input can be live the same day. You'll have the best tools available and no limits on what you use to get there. Outcomes are defined explicitly and measured weekly: