Full-stack Engineer, AI in London

About TORTUS AI

At TORTUS we work with clinicians and hospitals to provide clinically safe and reliable AI. We use speech-to-text and LLM models to infer information about a patient visit and provide verified documentation and actions — eliminating the documentation burden that costs clinicians hours every day. We've recorded over 1 million consultations. Clients include London Ambulance Service and a large proportion of UK GP surgeries. We are now entering our next phase: an AI doctor product and a ~£30M Series A raise.

About the role

Same generalist full-stack profile as our Full-stack Engineer role, but with a higher and more specific bar around AI integration. We are not looking for engineers who have called an LLM API. We are looking for engineers who have thought deeply about what it means to build reliably on top of probabilistic systems.

What you'll do

• Everything in the Full-stack Engineer role
• Design and build product features integrating deeply with LLMs in production
• Build defensively against hallucinations, output instability, and distribution shift
• Contribute to approaches for validating AI output quality in multilingual contexts — e.g. validating medical notes generated in a language other than English
• Work with our AI research team to implement architectures and enhancements they propose

What we're looking for

• All of the Full-stack Engineer profile, plus:
• Production experience building systems integrating with LLMs — well beyond API integration
• Strong intuition for failure modes of non-deterministic systems: output validation, fallback strategies, confidence thresholds, human-in-the-loop design
• Ability to reason about AI quality in adversarial or ambiguous conditions
• Familiarity with prompt engineering, RAG, fine-tuning, or LLM evaluation frameworks

Nice to have

• Experience in clinical AI or high-stakes AI domains
• Exposure to AI safety concepts applied in production
• Contributed to internal AI evaluation or red-teaming exercises

What does wild success look like?

In 12–18 months, you are the go-to person when the team hits a hard problem at the LLM integration layer. You have built evaluation pipelines that give the team confidence in what we ship. The Principal AI Researcher seeks you out as the translation layer between research ideas and production reality.

Where does this role take you?

This is one of the highest-leverage engineering roles at TORTUS as AI becomes central to everything we build. You could grow into a staff engineer role specialising in AI systems, or move toward technical leadership of the Lifeboat squad as it scales.

Our stack

GCP · GKE · Terraform · GitOps & ArgoCD · PostgreSQL + Redis · TypeScript across the stack (React, NestJS) · Python/FastAPI for ML workloads · Pub/Sub for events · Self-hosted ML models

What we offer

• 28 days holiday + bank holidays
• Latest MacBook and the right equipment
• Office-first culture based at Holborn Town Hall — typically Tuesday, Wednesday, Thursday in office; Monday and Friday flexible
• The chance to genuinely transform healthcare