Futurice's Senior AI Engineer role is a full-time position in the UK Tech team, with fully remote work across the UK and EU and a salary of £60,000 to £90,000 a year depending on experience. The role designs and builds AI applications, including MCP servers, voice-to-voice real-time applications, retrieval-augmented generation bots, and multi-agent systems. The UK office hub is London, while the position remains remote-first.
The senior AI engineer works with Python, RAG systems, agentic architectures, and at least one of Azure, AWS, or GCP. The role requires five or more years of hands‑on AI or applied data‑intensive software engineering experience, along with production software development, model deployment, and client communication. The position includes mentoring junior engineers, reviewing code, and contributing to shared AI tools and engineering practices. Futurice's interview process includes a talent screening call, a skills-based interview, a coding task, and a values interview.
Role: Senior AI Engineer
Employer
Futurice
Location
Fully remote across the UK and EU, with London as the UK office hub
Salary
£60,000 to £90,000 per year, depending on experience
Employment type
Full-time
AI Applications From Idea To Production
The role combines software delivery with AI engineering. That combination places the work within production AI work, where model development connects directly to software delivery. A Senior AI Engineer designs MCP servers, voice-to-voice real-time applications, RAG bots, and multi-agent systems, then deploys cloud-native generative AI applications into production. Client workshops and proposals turn technical possibilities into scoped proof‑of‑concept work, while fast demos give clients something to test before a larger build begins.
Futurice expects the engineer to work across ideation, prototyping, deployment, and maintenance rather than stop at model experimentation. Production-quality work includes clear interfaces, repeatable deployments, evaluation plans, and ownership after release. The position also contributes to shared AI tools, frameworks, and ways of working that support later projects.
Model selection, prompt design, customization, evaluation, deployment, and operations form the Google Cloud generative AI lifecycle. RAG connects a model to verifiable data sources, while vector embeddings and vector databases help retrieve relevant material for a prompt. Human review remains part of evaluation when model output affects quality, safety, or a customer‑facing service.
LLM, RAG, And Agentic Systems
The engineering work covers the full LLM stack, from training and fine‑tuning through application integration. The role requires practical understanding of LLM deployment, NLP, machine learning, data preprocessing, feature engineering, and model evaluation. Python forms the main programming requirement, with Pinecone, HuggingFace, LangChain or LangSmith, and CrewAI named as relevant tools.
Agentic architectures add tool use, orchestration, evaluation, and multi‑agent patterns to conventional model integration. A production system must handle the movement between model calls, external tools, retrieved information, and user‑facing outputs. The role also addresses data quality, bias mitigation, model interpretability, accuracy, efficiency, and reliability instead of treating the model as a finished component.
Model deployment, assurance, secure maintenance, and lifecycle management are core parts of the UK machine learning engineer capability framework. Those practices match the role's responsibility for production systems and its expectation that an engineer can maintain software after release.
Client Workshops And Engineering Review
Client work begins before implementation. The engineer advises clients on AI strategy through workshops and proposals, translates a business question into a technical approach, and prepares a proof of concept that can be tested quickly. Client‑facing communication matters because the role must explain model behaviour, system limits, trade‑offs, and next steps to people who may not work in software or machine learning.
The role includes code review and mentoring for junior engineers. Collaboration extends to application engineers, software engineers, product managers, and designers, with shared tools and frameworks carrying knowledge from one project into another. Python, model development, testing, risk assessment, deployment, and bug fixing are all part of the National Careers Service AI engineer profile.
Production Testing, Security, And Monitoring
AI systems need more than a successful prototype. Data quality work covers accuracy, completeness, permissions, security, representativeness, and update frequency. Separate training, validation, and test data support a clearer performance assessment, while unseen‑data testing helps expose overfitting and weak assumptions before release.
Production AI work includes security design, integration with existing services, scalable deployment, and monitoring against business objectives. A team also needs a way to detect incidents, identify when retraining is necessary, and evaluate the live system after changes. These controls form part of the GOV.UK implementation guidance for public‑sector AI projects, and they give the role's deployment work a concrete operational frame.
Futurice's Senior AI Engineer role applies those principles to client systems rather than isolated research experiments. The engineer evaluates model accuracy and reliability, considers data quality, bias, and maintains cloud‑native applications after they reach production. A clear test plan, monitored service, and documented operating approach reduce the gap between a convincing demo and a dependable application.
The UK Team And Remote Practice
Futurice was founded in 2000 and has more than 800 people across Europe. Its services include data, AI, design, strategy, and software engineering, with teams helping organisations build their own internal capabilities. The UK team includes designers, strategists, and software engineers, so the Senior AI Engineer works in a setting that combines technical delivery with design and business work.
Futurice's London office is the UK hub for a remote‑first team. Regular full‑team days bring colleagues together in person, while the role's day‑to‑day work remains remote across the stated UK and EU area. The company describes its working values as trust, transparency, care, and continuous improvement.
The Interview Stages And Role Package
The recruitment process has four defined stages, with each stage testing a different part of the role. The first conversation covers the position and the candidate's background. The skills interview examines practical judgement, the coding task tests implementation ability, and the values interview addresses collaboration and working style.
- Talent screening call: 30 minutes
- Skills‑based interview: 45 minutes
- Coding task: 1 hour
- Values interview: 45 minutes
Benefits
The package includes a personal learning budget of £1,200 per year, private health insurance through WPA, pension contributions of 6% or more, and £50 per month for wellbeing support through Juno. Annual leave starts at 25 days plus bank holidays and a birthday day off. One additional holiday day is added each year after three years of service, up to a 30-day limit.
Futurice also lists tiered parental leave, an electric vehicle salary sacrifice scheme depending on tenure, six‑month career reviews, mentoring, and knowledge‑sharing sessions. The listed package connects formal development with practical support for health, family leave, and long‑term progression.
Futurice's listed UK client work includes VW, Audi, RAC, B Corp, Ashurst, Moneycorp, Parkdean Resorts, East West Rail, Plan International, and WRAP. The company also lists generative AI work for KONE and a B Lab assessment platform that reached 70% automation with 95% accuracy.
The role's no‑sponsorship condition makes existing UK work eligibility part of the application decision.
#J-18808-Ljbffr
