AI Agentic Engineer - Build LLMs, Agents & RAG

Job Overview

We are a startup digital agency where every project, client, and day is unique and agentic‑first. In e‑commerce we build on Shopify, Liquid, Next.js, and custom AWS microservices that take Shopify beyond the defaults. On the marketing side we create highly customised Next.js and React sites with careful design, animation, and performance focus. In automotive and other industries we build full platforms, AI tools, and integrations—from stock‑car locators to configurators.

Across all projects we bring AI to the work through our own products. Nothing is fixed; you will work on long builds that run for months and on shorter spikes to explore how a business can use AI, rethinking how agency work is done.

Engineers who are excited by agentic and AI‑native systems will turn models into products that hold up under real users, load, and time. You’ll build LLM integrations, agents, and RAG pipelines, treat prompt engineering as a code discipline, and ensure systems are reliable through evaluation, guardrails, and solid testing. Whether you are building your craft or already shaping AI architecture for a project, we want to hear from you.

We hire across levels.

Building your engineering craft—curious, hands‑on, and eager to grow. You implement from clear briefs—LLM integrations, prompt scaffolding, RAG components—and write tests and evaluations that keep them honest. You use AI coding tools (Claude Code, Cursor, Copilot) under supervision, ask about purpose and constraints before you start, and your work is reviewed before merge.

Own an AI feature end‑to‑end—a agent flow, a RAG pipeline, an integration across its error paths and guardrails—delivered independently with light oversight. Build evaluation harnesses to prove quality, use AI coding tools autonomously in production, and contribute to the shared AI backbone (the AI‑brain MCP and skill library).

Own the AI outcome of a project—define the integration shape, multi‑agent architecture and orchestration, evaluation strategy and guardrails, and write ADRs that document why. Provide technical credibility, contribute patterns and skills back to the shared backbone, mentor other engineers, use AI to define new ways of working, and think long‑runtime as the AI landscape shifts.

• Experience building with LLMs or AI systems in production—integration, agents, RAG, or similar.
• Strong software engineering fundamentals in Python and/or Type Script, and confidence designing and working with APIs.
• Habit of treating prompt engineering as a code discipline with a real understanding of evaluation, guardrails, and reliability.
• Solid grasp of testing and quality: build systems that hold up under real use, not just demos.
• Genuine interest in agentic systems and AI landscape; judgement to apply it well.
• Openness to collaborating with cross‑functional teams—from designers to strategists—to clients.
• Enjoy working as part of a team, not a remote desk that prefers isolation.

• Languages & core:
  Python and/or Type Script, strong RESTful API design.
• AI & agents: LLM integration, RAG and vector databases, agent frameworks, multi‑agent orchestration, MCP and agent infrastructure, prompt engineering, context and memory management, evaluation and guardrails.
• Testing & quality: automated testing, evaluation harnesses, observability and monitoring of AI systems.
• Cloud & infrastructure: cloud platforms, containerisation with Docker and Kubernetes, CI/CD (infrastructure‑as‑code with Terraform a plus).
• AI tooling:
  Claude Code, Cursor, Copilot, Google Antigravity.

We are a bunch of colleagues originally from similar backgrounds who got tired of how things were usually done and set out to find better ways. That is the whole reason this company exists, and it is still what we are doing. We are not a place of rockstars and egos. People here ask questions out loud, show rough work without flinching, and would rather help each other get something right than look clever alone.

We push for continuous improvement and collaboration.