1 day ago
London, United KingdomStaff+
Responsibilities
- Architect, develop, deploy, and maintain production-ready AI systems, agentic workflows, multi-agent orchestration, RAG solutions, and intelligent automation.
- Design and integrate agent tools, APIs, and Model Context Protocol interfaces with enterprise systems, databases, and external services.
- Build compound AI architectures using large language models, specialised models, semantic search, classification, re-ranking, and deterministic workflows.
- Develop evaluation frameworks, automated testing, guardrails, and observability for model behavior, output quality, hallucinations, latency, and operational performance.
- Design backend services, APIs, and integration layers using Python, TypeScript, Node.js, and cloud technologies.
- Integrate AI capabilities into web applications and digital platforms with front-end, mobile, and product engineering teams.
- Manage token usage, caching, latency, rate limits, fallback mechanisms, and graceful degradation across AI services.
- Establish CI/CD, regression testing, version control, and reliable deployment practices for AI systems.
- Use cloud and model platforms including Azure OpenAI, AWS Bedrock, Google Cloud Vertex AI, OpenAI, and Anthropic.
- Provide technical leadership, architectural guidance, client communication, and solution-shaping across multidisciplinary teams.
Requirements
- Significant experience in backend, platform, or software engineering, including distributed systems, microservices, asynchronous processing, API architecture, and cloud infrastructure.
- Proven hands-on experience designing, deploying, and maintaining generative AI or agentic systems in production.
- Advanced proficiency in Python and/or TypeScript/Node.js and experience writing clean, testable, maintainable production code.
- Practical understanding of LLM architectures, agentic workflows, tool calling, RAG, embeddings, vector databases, and hybrid retrieval.
- Experience with agent orchestration frameworks such as LangGraph, CrewAI, AutoGen, Vercel AI SDK, or custom orchestration solutions.
- Experience developing AI evaluation and observability capabilities using tools such as LangSmith, Arize Phoenix, Braintrust, Helicone, or OpenTelemetry.
- Strong understanding of software architecture, security, authentication, data management, enterprise integration patterns, and cloud infrastructure.
- Experience implementing automated unit, integration, and regression testing and evaluating non-deterministic AI systems.
- Understanding of modern digital product development, web applications, headless CMS platforms, APIs, and front-end technologies.
- Ability to balance cost, latency, performance, security, maintainability, and commercial objectives.
- Experience communicating complex technical concepts to clients and non-technical stakeholders and providing architectural leadership.
- Experience in a digital agency, technology consultancy, or product-led environment is expected.
- Experience fine-tuning smaller or open-weight models such as Llama or Mistral is advantageous.
- Experience with design-to-code workflows, Figma integrations, design systems, or generative user interfaces is beneficial.
Benefits
- Extensive benefits are available through AKQA.
- The role follows a hybrid approach with required in-office days at the London studio.
- Workplace accommodations or flexibility can be discussed with the hiring team.
Tech Stack
Categories
About AKQA
AKQA is a design and innovation agency that builds digital products, services, and brand experiences for consumer and enterprise clients across sectors such as retail, automotive, and finance. The company operates on a services and consulting model spanning strategy, design, engineering, data, and media. Founded in 1994 and headquartered in London, AKQA is part of WPP and maintains studios across dozens of cities worldwide.