AI Software Engineer
Trtl is a travel accessories brand based in Glasgow City Centre, Scotland. With over two million customers worldwide, we make travel easier — starting with our hero product, the Trtl Travel Pillow, reinvented from the ground up to solve a problem no one had properly fixed.
We are a small, ambitious company that punches well above its weight — around £30M in annual revenue, selling across Amazon, Shopify/DTC, and TikTok Shop in the UK, US, and internationally. We have a clear ambition to scale to £50M+ and AI is central to how we get there.
We have been building AI-powered tools and workflows into the business for two years. AI is already core to how we operate. This role exists to accelerate that — materially and measurably.
We are looking for a Mid-Level AI Software Engineer — someone with strong software development foundations and growing hands-on experience with agentic AI systems.
You will work directly within our technology team, reporting to our Head of AI. You will help build, extend, and maintain a growing ecosystem of AI-powered tools and integrations already live across the business — working closely with colleagues in marketing, operations, customer support, and finance to turn real problems into working solutions.
We do not expect you to have shipped production-grade agentic systems at scale. We do expect you to understand how they work, to have built with LLMs in earnest — whether in a commercial setting or your own projects — and to be ready to grow fast in an environment that takes this seriously.
If you want to learn what AI-native commerce looks like in practice and have the engineering foundations to contribute from day one, this role is for you.
1. Build & Maintain AI-Powered Tools
Integrate AI models: Work with large language models to power tools — classification, summarisation, structured generation, and decision support. We use Claude (Anthropic) as our primary model.
3. Collaboration & Problem Solving
Strong practical understanding of LLM prompting: system prompt design, chain-of-thought, structured outputs, prompt versioning
Aware of common failure modes in agentic systems: prompt injection, tool misuse, hallucinated tool calls, context window limits
Experience building and maintaining microservices architectures, including inter-service communication, data pipeline design, and API integration
Hands-on experience building agentic AI systems — not just calling LLM APIs, but designing multi-step agents, handling tool use, managing context, and reasoning about failure modes. Personal projects count.
Trtl (Coolside Limited) · Glasgow, Scotland · trtl.com · "Never a passenger — always a wide awake explorer"
