AI-Native Full-Stack Engineer

About R2 Solutions AI

R2 Solutions AI is a small consulting firm that designs and delivers AI-powered systems for mid-market clients: agentic architectures, Salesforce development, Azure data infrastructure, and the software that ties them together. We run lean. Every engineer owns engagements end to end and talks directly to the people who will use what they build.

We build software the way we think it should be built in 2026: engineers define the problem and the proof; coding agents do the typing. We are hiring our next engineer to do exactly that.

The role

This is not a traditional coding job. Your output is a precise, testable specification and the agent configuration that turns it into production code. The agent's output is the code. Your judgment is what makes the difference between a demo and a system a client runs their business on.

Expect your time to split roughly like this:

1. 15–20%

   Discovery with clients and stakeholders: existing systems, data, constraints, what "done" actually means.

2. 40–50%

   Turning that understanding into agent-readable specs, architecture decisions, test plans, and the skills, hooks, and CLAUDE.md files that make agents produce tested code the first time.

3. 35–45%

   Orchestrating and supervising agent implementation, reviewing and hardening generated code, and shipping through CI/CD.

You will work on a team of one to two engineers per engagement, often as the technical lead on that engagement, with a principal available for architecture review and client escalation. You set your own daily plan against the engagement goals.

This is a contract-to-hire role. The initial engagement is a 6-month, roughly full-time (35–40 hrs/week) independent contract at an hourly rate. At the end of that period, if it is working for both sides, we convert you to a full-time principal-track role. We treat the contract period as a mutual trial, not a holding pattern: you get real engagements and real ownership from week one.

How we work

Spec-driven.

Every feature starts as a spec with acceptance criteria and test cases. Agents implement against the spec, not against a conversation.

Agent loops.

Spec → agent loops → validation → deploy. Short cycles, frequent merges, nothing lands without passing the harness.

TDD as a gate.

Red-green-refactor is enforced, not aspirational. Tests gate agent output; a failing test is the agent's to-do list.

CI/CD as the harness.

The pipeline is the harness. Linting, unit, integration, and contract tests run on every push. Deploys are automated.

Skills and hooks as design.

If you repeat an instruction to an agent three times, you turn it into a skill. If the agent keeps making the same mistake, you write a hook. The agent environment is a first-class deliverable.

Review is a skill.

Generated code is reviewed to the same standard as hand-written code, with extra attention to security, error handling, and the things agents are confidently wrong about.

What you will do

• Lead technical discovery with client stakeholders and translate business requirements into architecture and executable specifications.
• Author and maintain agent configurations (skills, subagents, hooks, MCP servers, CLAUDE.md) that encode project conventions, testing standards, and domain knowledge.
• Drive implementation through Claude Code or equivalent agents across the full stack: APIs, data models, integrations, and front ends.
• Design and own the test strategy and CI/CD pipeline for each engagement, from first commit to production.
• Review, test, and harden agent-generated code to production quality; own what ships.
• Present progress, trade-offs, and recommendations directly to clients; manage scope and expectations.
• Grow into solution architecture: own system design across engagements, evaluate platforms, and shape R2's delivery practice.

What we are looking for

• 4+ years of professional software engineering experience with hands-on full-stack delivery (TypeScript/JavaScript or Python; React or similar; SQL and data modeling; cloud deployment on Azure or AWS).
• Daily, hands-on use of AI coding agents (Claude Code, Codex, Cursor, or similar) as your primary interface for design, implementation, testing, and refactoring. If agents have not materially changed how you work, this role is not a fit.
• Experience writing specifications that agents can execute against, and a clear point of view on what makes a spec good.
• Experience authoring agent customizations: skills, subagents, hooks, MCP servers, or equivalent. You should be able to show us one.
• Test-driven development as a working habit across unit, integration, and contract layers, including using tests to constrain and verify agent output.
• Experience building and maintaining CI/CD pipelines (GitHub Actions, Azure DevOps, or similar) and treating them as the quality gate.
• Comfort working directly with non-technical stakeholders: eliciting requirements, explaining trade-offs, saying no well.
• Demonstrated ability to self-direct on a small team or solo: you decide what to do next, and you are usually right.

Nice to have

• Salesforce platform development (Apex, LWC, Flow, Agentforce) or other enterprise SaaS platform work.
• Azure data stack: Data Factory, Azure SQL, Databricks, or equivalent.
• Experience building LLM-powered features: tool calling, RAG, evaluation harnesses, cost and latency control.
• Consulting or client-facing delivery experience.

How we hire

We screen for how you actually work, not how you interview.

1. 1

   Intro call.

   A 30-minute conversation about your current agent workflow. Bring a real example.

2. 2

   Take-home.

   Ship a small feature using spec-driven development and an agent. You submit the spec, the skill or hook files, the test suite, and the code. We care about the first three more than the fourth. Budget 3–4 hours; we pay for your time.

3. 3

   Working session.

   Walk us through your take-home. Then we hand you a vague client ask and watch you turn it into a spec, live.

4. 4

   Final conversation.

   With a principal. References, then a contract offer.

Compensation and logistics

Conversion:

Full-time salary and bonus structure are set at conversion based on demonstrated performance during the contract period; we will share the target range during the interview process.

Benefits and equipment:

As an independent contractor you carry your own benefits, insurance, equipment, and taxes. We provide client-environment access and the AI tooling licenses required for engagement work.

Location:

Remote within the US; occasional client travel, billed at cost.