Joan Alemany

End-to-end development of a LegalTech platform, managing both Cloud architecture and the implementation of Artificial Intelligence-based systems. Ingestion Pipeline: Redesigned document ingestion from manual PDF uploads to automated scripts against official APIs (BOE), with optimized chunking, enriched metadata, deduplication, and classification. RAG Optimization: Reduced response times through vector search indexes, reranking, and tuning the k parameter. Migrated LLM invocation from Lambda to the backend for greater control and efficiency. Cloud Infrastructure: Separated dev/prod environments into isolated instances provisioned with Terraform. 30% reduction in AWS costs. Security and Availability: Implemented AWS Cognito, hardened infrastructure security, and resolved production outages during peak usage. CI/CD: Evolution from manual deploys to Docker containers with automated pipelines via GitHub Actions. Full Stack: Refactoring of monolithic React components, creation of backend APIs with FastAPI, and end-to-end integration (payment gateways, notifications, calendars).

Refactored legacy code in Next.js, improving loading performance and maintainability. Developed React/TypeScript components for projects for brands like Coca-Cola and Jijona. Implemented custom interactive solutions (mini-games, with

Developed and implemented user interface components with React and TypeScript for high-profile client projects. Optimized and refactored existing code in Next.js, improving performance and maintainability. Implemented a user authentication system using PHP. Contributed to content management, performing article updates in the Sitecore CMS.