Platform engineer for AI agents

The Kite Factory; Greater London; On-site; GBP 70,000 - 110,000; Full time

Benefits offered by this job

Competitive salary

Training & development

Job summary

The Kite Factory в Лондоне ищет специалиста по платформе и DevOps, который построит и запустит платформу для агентов: оркестрацию, маршрутизацию моделей, контроль доступа и аудит. Вы будете работать с Head of Product и закладывать фундамент для устойчивой AI трансформации.

Требуется опыт в Azure, Terraform/Bicep, CI/CD, PostgreSQL, работе с LLM в продакшн, знание Python или R. Кандидат должен быть креативным, проактивным и прагматичным; опыт в медиа/рекламе будет плюсом.

Qualifications

• Опыт разработки и сопровождения платформ с управлением доступом и аудитом.
• Опыт работы с продакшн-LLM системами, оркестрацией и мониторингом.
• Опыт DevOps, инфраструктуры и CI/CD в реальных проектах.

Responsibilities

• Построение и запуск платформы для агентов: оркестрация, маршрутизация, очереди и окружения.
• Обеспечение контроля доступа и аудита, least privilege и аудит логов.
• Настройка и Stand up метрик затрат (cost metering) и просмотр нагрузки.

Skills

Proactivity

Creativity

Pragmatism

Boldness

Problem-solving

Communication

Tools

Azure

Terraform

Bicep

PostgreSQL

CI/CD

Python

Job description

Описание

This is a new role in our product function working with the Head of Product, you will build and run the platform that hosts our AI agents: orchestration, model routing, permissions, audit logging and cost metering and the infrastructure behind our evaluation and quality-gate system. The remit focusses on setting the groundwork for sustainable AI transformation

Задачи

Build and run the agent platform: orchestration, model routing, queues, retries, environments and deployments Own access control for humans and agents alike least privilege by default, quarterly reviews, no orphaned credentials Put audit logging on every automated action and keep the logs queryable, attributable and auditor-friendly Stand up cost metering and alerting across model API usage, GPU workloads and vendor licences, with unit costs visible per workflow Run the evaluation infrastructure behind our quality gates in partnership with the data & analytics team Manage vendor integrations - model providers, RPA tooling and legacy media systems - and cost their brittleness honestly rather than wishing it away Harden and document the agency’s infrastructure: runbooks, degraded modes, recovery procedures, secrets hygiene Support the wider transformation programme: every team is adopting agent workflows on a paced schedule

Требования

Major cloud in-production: We use Azure with infrastructure as code (Terraform, Bicep or equivalent) and CI/CD R or Python as primary everyday language. PostgreSQL in-production: Comfortable around data pipelines and warehouses. Experience running LLM-backed systems in production, gateways, orchestration frameworks, observability. Deployment and testing open weight models. 3+ years' experience in platform, DevOps, infrastructure or similar engineering roles We are looking for the following traits from someone in this role; passion, creativity, problem-solving, boldness, proactivity, practicality and pragmatism. Будет плюсом: PostGis and Frontend UX design experience advantageous but not essential; Ideally work experience working in media, marketing or advertising

Условия

we offer a competitive salary, a range of benefits, a huge focus on training and development with dedicated training time, partnerships with industry specialists, and the opportunity to progress and develop your career with the support of a fantastic team.