Owned the implementation of a major refactor to a core multi-tenant sign-up decision engine, applying design patterns to make routing rules extensible; fixed production defects causing duplicate applications and blocked retries with test-first regression coverage. Built a shared offline-conversion batch pipeline for sales and qualified-lead events across Google, Meta, and Microsoft Ads, including privacy-preserving customer matching with normalized, SHA-256-hashed email and phone data. Optimized a core customer-data endpoint, cutting response time from 30s to under 1s (over 30x) by using EXPLAIN ANALYZE and correlated LATERAL joins to eliminate global aggregations over 2M+ rows.

Architected and built the backend orchestration layer for its ML platform, integrating dynamic model schemas, validated S3 training-data ingestion, real-time inference, and end-to-end prediction traceability; supported more than 22 million disease predictions across Latin American healthcare institutions. Led the technical and documentation effort to achieve ISO 27001 certification and sustain compliance for the following year, implementing AWS encryption, network segmentation, and secure backups. Architected a serverless AWS pipeline processing more than 500,000 inferences per batch using SQS, Step Functions, Lambda, S3, and RDS. Built and evolved production-oriented LLM workflows using OpenAI within a multi-role medical research platform, integrating agent orchestration, streaming interfaces, asynchronous processing, persistent APIs, and server-side security.

Built and maintained web and mobile applications for consulting clients using React.js and React Native.

Led a 7-person team delivering four software development and support projects for Volkswagen, Audi, and Autotek, owning estimation, planning, execution, and client delivery.