Full-Stack AI Engineer (AI-Native)

Our client is building AI-powered products where autonomous agents perform meaningful work under human supervision. They are looking for an experienced full-stack engineer who treats AI agents as core development tools, not occasional assistants.

The role

Small engineering teams can now ship what once required dozens of people. The engineers working at this pace do more than write code. They create precise specifications, direct coding agents, build systems that keep them reliable, and validate the final product with real users.

The code may be generated. The engineer still owns the specification, architecture, testing, and outcome.

You will join a highly autonomous team that already works this way and take ownership of features from initial concept through production.

What you’ll build

Our client is developing two AI-native products:

• An AI workforce platform where specialist agents complete real client work through Slack, email, and chat. Outputs pass through automated safeguards, evaluations, and human expert review before reaching the client.

• An AI-native recruitment product built around the same combination of autonomous agents and accountable human oversight.

You will work across both products and own a defined area end to end.

What you’ll do

• Build and ship complete features across a TypeScript and Node.js API, Python agent runtime, and Next.js frontend.

• Take products from early wireframes and technical specifications through testing and production deployment.

• Design agent workflows that incorporate human review, escalation, corrections, and feedback loops.

• Build drafting pipelines, review queues, and quality controls for high-stakes AI output.

• Create evaluation sets that gate CI and prevent regressions.

• Turn real incidents and failure modes into guardrails and automated tests.

• Improve prompt assembly, tool calling, context management, MCP servers, and the internal knowledge layer.

• Monitor and optimize model latency, reliability, and cost.

• Question requirements, identify weak assumptions, and propose better solutions when appropriate.

• Demonstrate working features directly to the founders each week.

What we’re looking for

• At least five years of experience shipping production software across the stack.

• Strong TypeScript experience and the ability to work confidently in Python.

• Experience delivering at least one LLM-powered product used by real customers.

• Practical experience with evaluations, guardrails, observability, and production failure modes. Prototype-only or demo experience is not sufficient.

• Strong product and interface judgment, with the ability to explain design decisions clearly across product and engineering teams.

• A working understanding of how modern AI development systems operate beneath the interface, including context assembly, agent loops, tool calls, structured outputs, and common failure modes.

• Fluency directing coding agents from the terminal.

• Experience customizing AI development workflows through skills, hooks, MCP servers, prompts, or purpose-built tooling is a strong advantage.

• The judgment to ask the right clarifying question before implementation begins.

• Direct communication and the ability to explain technical tradeoffs in plain language.

• A public record of learning or building, such as meaningful GitHub projects, open-source tools, technical writing, or shared experiments.

How the team works

• AI agents produce much of the initial code. Engineers own the specifications, prompts, safeguards, evaluations, and final decisions.

• Planning comes before implementation. A clear plan is valued more than a fast but incorrect result.

• Automated review gates pull requests and supports rapid iteration.

• Engineers can deploy freely to development environments, while production releases are managed through pull requests.

• The team demonstrates working software to the founders every week.

• Work is remote, asynchronous, and intentionally light on meetings.

• Requirements and architectural decisions are open to challenge.

• Intellectual honesty matters. Saying “I don’t know yet, but here is how I would find out” is preferred to an unsupported answer.

What’s offered

• Full ownership and autonomy over your area of the product.

• Direct access to the founders, with minimal organizational layers.

• A dedicated budget for AI development tools.

• Remote and async-friendly working.

• A fast-moving environment with weekly production releases.

• The opportunity to shape how an AI-native engineering organization builds software.

About our client

Our client is building systems that allow AI agents to act on behalf of real, accountable people.

Its products combine capable autonomous agents with human judgment. AI handles the repetitive and scalable work, while human experts review important decisions and remain accountable for what reaches the customer.

The company is interested in a more useful question than whether AI will replace engineers: what can one exceptional person accomplish with a hundred tireless collaborators?

If that is the kind of engineering future you want to help create, we would like to hear from you.