Senior Engineer (Agentic Developer) - #2157725

Reed.ai is looking for a Senior Engineer (Agentic Developer) to join their team in Holborn, London.

Join a collaborative, forward-thinking team that values creativity, experimentation, and impact—and help redefine the future of recruitment through intelligent, human-centered technology.

Reed.ai is building its recruitment platform using an AI-agent-first way of working. Coding agents do most of the typing; engineers are responsible for what gets built, why, and whether it is good enough to deploy. The platform was built this way from the start with agent fleets working in evidence-gated waves under human direction and it has now launched its product to the market. You'll join a startup-style venture backed by the strength of an established business, at the moment delivery is shifting from a contracted build team to a permanent in-house core: the ways of working you help set now are the ones the platform scales on.

We're looking for a senior engineer who has made the shift already. You will turn product intent into clear specifications. You will direct and coordinate coding agents across the stack. You will design the tests, evaluations and review gates that keep quality high. And you will be fully accountable for every line that reaches production, whoever, or whatever, wrote it.

You'll write less code than in a traditional senior role and read far more. Success is measured by the outcomes you deliver, how reliable and safe the platform is, and how much the team's agentic workflow improves over time. Lines committed is not a measure we use.

• Spec-first. Product intent arrives as a brief; work starts by writing the implementable specification and playing it back in plain language before build. The specification is the means of production, its precision sets the ceiling on what the agents can safely build.
• Evidence-gated. Work ships when its gate proves green, tests passing on enforcing infrastructure, invariants held, migrations verified — not when someone asserts it. Work that cannot prove its gate reverts itself; a clean tree is a passing outcome.
• Discernment. AI output is a claim, not a fact, a draft until a named human can defend its content without the tool. Fluent and plausible is not the same as correct, and passing output onward at face value is the failure mode we guard against hardest.
• Drift-aware. The codebase moves faster than any document describing it. Verifying against the live repository is step one of every task; divergence between the signed-off spec and the code is a defect, even when the drifted version is nicer.
• Communicating constantly. Shared surfaces are claimed before they're touched, coordination notes travel inside pull requests, and the daily standup ends with “what did I change that someone else's brief still assumes?”

• Translate product requirements into testable specifications and tasks, then plan and supervise coding agents across the full stack.
• Maintain context packs, repository guidance and decision records to support consistent agent output, writing code directly where needed.
• Define acceptance criteria and independent tests before generation, keeping security controls enabled throughout verification.
• Review agent-generated code for correctness, security and architectural fit, and maintain automated quality gates in CI/CD.
• Evaluate AI features for matching accuracy, regressions, adversarial cases, bias and fairness across candidate groups.
• Maintain coherent architecture, restrict agent access and uphold security, privacy, UK GDPR and applicable recruitment AI obligations.
• Improve agentic workflows, assess new tools and models, and mentor engineers in spec-first, agent-driven development.
• Collaborate with product, design and customer experience to test hypotheses and deliver weekly releases behind feature flags.

• Substantial commercial engineering experience at senior or lead level, building and running production SaaS systems.
• Hands-on delivery using AI coding agents such as Claude Code, Cursor or Codex, with evidence of effective workflows, failures caught and results achieved.
• Strong critical code-review skills and clear written communication to produce precise specifications for agents.
• Strong PostgreSQL, relational design, REST API and cloud-native fundamentals (AWS preferred), with depth in at least two of Go, Python and TypeScript.
• Experience in automated testing, CI/CD, DevOps and live database migrations, alongside a working understanding of security, data protection and GDPR.
• Desirable: experience building or evaluating LLM features, including retrieval, embeddings, vector search, agent orchestration and evaluation frameworks.
• Desirable: experience with OpenSearch/Elasticsearch, Docker, Kubernetes or event-driven architecture, plus front-end performance and accessibility.
• Desirable: recruitment, HR technology or matching-system knowledge, and experience in startups or new ventures within larger organisations.

• Hybrid working as standard
• 25 days annual leave plus bank holidays
• Flexible holiday scheme
• Paid time off to move home
• Contributory pension scheme
• Enhanced family leave benefits
• Insurance benefits including life assurance
• Love Mondays events
• Discount scheme including gyms and popular retailers
• Range of wellbeing and mental health support avenues
• Office in a fantastic location, with countless bars, restaurants and theatres right on the doorstep.