Staff AI Engineer, Developer Productivity

Description

The company makes a ring and connected app that give members daily insights into readiness, activity and sleep quality. It is growing quickly and serves millions of people.

The AI Transformation team is changing how software gets built. Every engineer has agentic tools, and the company already runs a pipeline that turns a well-specified ticket into a reviewed, tested stack of commits. That pipeline runs interactively, headless and on a fleet of cloud sandboxes, with its own evals and tracing.

The next goal is to cover the full delivery loop, from spec to operations, so hundreds of engineers across iOS, Android, backend and AI services can ship this way every day. The team is LLM-forward and moves fast, favouring rapid prototyping and iterative loops.

Success means validated changes reach members sooner and more safely. It is tracked through time to a trustworthy signal, escaped defects, cost per merged change, and how many squads choose the paved path because it works better.

Responsibilities

• You do not need to cover all of these on day one. Own the technical direction of the existing ticket-to-PR pipeline, including its architecture, agent orchestration, context handling and tool calls. Raise its reliability, throughput, cost and developer experience across mobile, backend and web repositories, and run it unattended wherever risk allows.
• Extend agent work past code generation into spec quality checks, test generation, automated review, CI failure triage and repair, release checks and post-deploy signals, so the system learns from what happens after merge.
• Build quickly: get a new workflow working within days, validate it against actual tickets alongside squads, then promote the parts that prove out into the production stack.
• Operate AI in production: build eval sets, LLM-as-a-judge grading, versioning, tracing and dashboards that show whether a workflow change helped, including a change that passes CI but is still wrong. Track merge rate, rework, time to signal, cost per change and escaped defects, and show them to both engineers and leadership.
• Design sandboxing, permissions, human checkpoints and guardrails scaled to risk, with privacy built in from the start. Work touching health data, security or infrastructure gets tighter controls, while work that is low risk and simple to verify can run without supervision.
• Model AI-assisted engineering standards across the organization through your own output, and help the engineers around you improve.
• Work with squads and their AI DRIs to find where the pipeline fails on real work. Fix the system rather than one output, and turn what works into paved paths, skills, templates and docs.
• Partner with platform, developer productivity, security and privacy, AI infrastructure, mobile and backend squads, product, design and engineering leadership. Make trade-offs explicit and keep everyone on one system instead of several parallel ones.

Requirements

• You have built and run production systems that other engineers depend on, such as developer tooling, internal platforms or AI infrastructure at scale. Demonstrated results count for more than years of experience.
• You have built and operated agentic coding workflows or autonomous loops in a real engineering organization, not only demos: background agents opening PRs, automated review, CI repair, test generation or similar, with adoption and measured results.
• You think about specification, review, testing, CI/CD, release and operations as one system, and you know where the bottleneck shifts as code generation gets cheaper.
• You set direction across teams you do not manage, write designs others build on, and make the hard calls on architecture and risk.
• You are strong in fundamentals such as mobile domain or backend APIs, concurrency, distributed systems, testing, debugging and reliability. You treat AI output as a draft and can show how you verified it.
• You are fluent in one of backend, iOS or Android and can work in the others. You use available tools to move across platforms and pick up new stacks quickly.
• You understand the responsibility of handling sensitive personal data and of giving agents access to real systems, and you design for both from the start.
• You thrive in ambiguity and shape requirements yourself through experimentation.
• You bring teams along, take feedback from skeptical engineers, share credit, and keep going when experiments fail or tooling changes. You communicate clearly with engineers, managers and non-engineering partners.
• Would be a benefit: depth with coding agents, including skills, MCP servers, hooks, sub-agent orchestration and headless runs.
• Would be a benefit: evaluation harnesses or benchmarks for coding agents, or LLM-as-judge grading tied to build and test results.
• Would be a benefit: sandboxed execution at scale, such as ephemeral dev environments, VM or container fleets, or macOS build infrastructure.
• Would be a benefit: background in developer productivity measurement (DORA, SPACE, DX Core 4) and in turning metrics into decisions without making them targets.
• Would be a benefit: experience shipping mobile apps or other products with slow, costly feedback loops.
• Would be a benefit: experience in digital health, wearables, MedTech or other privacy-sensitive or regulated products.
• Would be a benefit: contributions to open-source AI projects, or a portfolio of projects built on the latest agent capabilities.
• Would be a benefit: experience mentoring engineers or leading an engineering community of practice.

Conditions

Hybrid work, with time spent at the London office in the UK.

You must already have the right to work in the UK. The company cannot sponsor a visa for this role.

The company is an equal opportunity employer, considers all candidates regardless of protected characteristics, and does not tolerate discrimination or harassment.

Benefits

• Competitive salary.
• A ring of your own, plus employee discounts for friends and family.
• Flexible working hours.
• A collaborative group of coworkers.