Deloitte
Madrid, Spain
Oct. 2023 Present
-
Senior AI Cloud Engineer
Sep. 2025 — Present
Work directly with clients to turn business needs into AI solutions, combining technical planning, architecture, and hands-on development while guiding teams in AI adoption.
Explore impact Back to overview — Senior AI Cloud Engineer, Sep. 2025 — Present
Selected impact
- Led the development of an agent factory using AWS Strands Agents and Amazon Bedrock AgentCore, building reusable blueprints with MCP and RAG integrations, enabling conversational deployment across AWS accounts.
- Worked directly with clients to translate business challenges into actionable technical proposals, defining functional and non-functional requirements, breaking down implementation tasks, and establishing delivery roadmaps and timelines.
- Built a production conversational RAG application for a client using Python, FastAPI, LangChain, and Amazon Bedrock, with OpenSearch retrieval and automated knowledge base synchronization, raising customer satisfaction by 150%.
- Led the corporate-wide rollout of Claude Code for a major client in Switzerland, traveling on-site to establish development best practices and train over 200 engineers, driving org-wide adoption of AI-assisted development practices.
- Built a Claude Code development harness integrating project context, specs, skills, and agents, increasing capacity-adjusted story points delivered per sprint by 40% compared with the pre-adoption baseline.
- Built and deployed an agent in a Teams channel using Python and Claude Agent SDK, integrating Confluence via MCP to deliver source-backed answers, with execution limits and application logging, reducing time-to-answer from three days to one.
-
AI Cloud Engineer
Oct. 2023 — Aug. 2025
Developed AI applications and cloud infrastructure across conversational AI, document automation, and predictive ML, from architecture design to production deployment.
Explore impact Back to overview — AI Cloud Engineer, Oct. 2023 — Aug. 2025
Selected impact
- Developed a production AI interview simulator using Java and Amazon Bedrock, extracting structured information from CVs and job descriptions to generate adaptive questions, detect contradictions, and deliver personalized feedback, using Amazon Transcribe/Polly (STT/TTS) for voice integration and DynamoDB for conversation persistence.
- Implemented a document automation system for an insurance client using Python, Amazon Textract, Bedrock and LangGraph, extracting invoice and policy data into DynamoDB and automating report generation, reducing completion time by 72%.
- Defined reference architectures for microservices, serverless, and Generative AI applications within the Cloud Center of Excellence, reviewing AWS architectures, deployments, and security against Well-Architected principles, with architecture and optimization recommendations reducing cloud costs by an average of 7% across reviewed applications.
- Built and deployed fraud detection models for a banking client using Python, XGBoost, CatBoost, and AWS SageMaker, enabling real-time inference through SageMaker Endpoints and improving fraud detection precision by 6%.
- Built GitHub Actions CI/CD pipelines to automate testing, Docker image builds and publishing to Amazon ECR, and Kubernetes deployments on Amazon EKS, with commit-based image tagging and approval gates for production releases.
