AI Platform Architect, Madrid

Stack

Responsibilities

• ▹Design and evolve the architecture of the corporate AI platform across hybrid environments: multicloud, on-premise and Mainframe
• ▹Build self-service capabilities so different technology teams can consume AI models easily
• ▹Define the model access and mediation layer through an AI Gateway, including credential management, routing, quotas and cost control
• ▹Design architectures for agentic systems: orchestration, context management, tool integration via MCP and function calling, and RAG patterns
• ▹Design the agent execution environment (harness), with context, tools, isolation, permissions and verification that produce repeatable, reviewable results
• ▹Define the lifecycle governance of models: selection criteria, approved models, quality evaluation and usage traceability
• ▹Add AI-specific security controls, especially against prompt injection, data leakage or improper access during agent execution
• ▹Define observability, traceability and cost-control mechanisms for the use of models and agents
• ▹Integrate AI capabilities into the secure software development lifecycle, including CI/CD, infrastructure as code and running on Kubernetes
• ▹Design the controls needed to ensure auditable AI use aligned with the banking regulatory environment
• ▹Take an active part in building the platform, making technical decisions and keeping the implementation consistent with the defined architecture
• ▹Document architecture decisions and components so that different technology teams can understand and adopt them

Requirements

• ▹Consolidated experience in technical architecture, usually 5 years or more in your area of specialization
• ▹Deep experience in at least one of these environments: cloud (Azure, AWS or GCP) or on-premise (on-premise Kubernetes or VMware)
• ▹Good understanding of hybrid architectures and integration between different technology environments
• ▹Experience designing or implementing self-service AI platforms and mediation mechanisms for model access
• ▹Solid knowledge of agentic system architectures, orchestration, context management, MCP, function calling and RAG
• ▹Experience or advanced knowledge of harness engineering and designing secure, controlled execution environments for agents
• ▹Knowledge of AI model lifecycle governance, evaluation, selection and traceability
• ▹Experience applying AI-specific security principles
• ▹Knowledge of observability and mechanisms to control and optimize the cost of models and agents
• ▹Experience with CI/CD, IaC and Kubernetes within the development lifecycle
• ▹Ability to design and document architectures and solve highly complex technical problems
• ▹Ability to work with still-evolving technologies, designing solutions that reduce coupling to specific vendors or models
• ▹Judgment to decide which processes can be automated with AI and which must keep human oversight or intervention depending on risk
• ▹Understanding of the banking regulatory environment as it relates to AI architecture and use
• ▹Product orientation: seeing the platform as an internal product that must answer the needs of real users and teams

Soft skills

What we offer

• ▹Permanent contract
• ▹3 days remote + 2 days in Alcalá de Henares
• ▹Internal training and access to certifications
• ▹Flexible compensation plan (health insurance, transport, childcare and meal vouchers)
• ▹Referral program: a bonus for recommending friends
• ▹Events, meetups, tech days, talks
• ▹26 days of rest (22 vacation days, 2 days of free disposal, and December 24 and 31 off)
• ▹Flexible hours: 8:30 to 18:00 Monday to Thursday and Friday 8:00 to 15:00; summer intensive schedule in July and August, 8:00 to 15:00

About the company

knowmad mood is a leading, constantly evolving digital transformation company with more than 3,000 creative, digital and innovative people. Its values are collaboration, innovation, commitment, fun and trust.