Nicolás Calvache Jiménez

Tech Lead & Software Engineer

• Architected and deployed Vertex AI voicebots for the Asturias Health Service (SESPA), automating 24/7 appointment management and handling 100,000+ calls per month.
• Deployed production outbound voice agents for Spanish occupational health insurers (Mutuas), orchestrating up to 60,000 calls per month and routing contacted patients into in-person or nursing appointments. Achieved patient contact rates of up to 70%, accelerating internal sick-leave case management.
• Built context-aware Generative AI and LLM agents for complex administrative flows within the Rovozia SaaS platform.
• Engineered Node.js REST APIs and real-time BigQuery and Firebase data pipelines for high-volume concurrent AI agents.
• Managed Google Cloud infrastructure with Terraform and automated CI/CD through GitHub Actions and Cloud Build.
• Delivered Grafana and Looker Studio observability, legacy-system integrations, 99% service availability, and GDPR-aligned data protection controls.

• Designed and deployed a serverless Google Cloud architecture using Cloud Run, Cloud Storage, and MongoDB Atlas, reducing infrastructure costs to near zero.
• Delivered the end-to-end deployment and backend integration of a React Native mobile MVP.

• Designed and developed an 'Advanced Voicemail' module for the existing Rovozia SaaS platform, leveraging Dialogflow and Node.js to enhance core telephony capabilities and user experience.
• Seamlessly integrated these AI-driven microservices into the legacy telephony architecture, significantly boosting operational efficiency by automating the handling of 300 to 700 customer calls daily.

• Co-led an Agile team of 6 interns, managing sprint priorities, task delivery, and code reviews.
• Modernized LMS modules using PHP, Mustache, and JavaScript while translating stakeholder requirements into technical deliverables.

• Contributed to the European EdTech initiatives LOVEDISTANCE and KOULU.
• Built REST APIs for Moodle integrations and improved internal interfaces using HTML5, CSS, and JavaScript.