Projects

Conversational AI agent built with the Anthropic SDK's native agentic loop — no orchestration frameworks. Runs live on this portfolio with multi-tier identity, persistent memory, Google Calendar/Gmail tools, Telegram integration, and a resilient multi-provider LLM fallback chain.

Production conversational agent that analyzes diagnostic studies through WhatsApp and routes patients to the appropriate specialist—surgeon or physical therapist—while keeping the physician as the final clinical decision-maker. Includes the proposed V2 evolution toward a hybrid pipeline with calibrated machine learning and an abstention layer.

At Heineken Mexico, I developed an ML model that predicts the probability customers will stop puchasing (churn). Separately, here I highlight another ML model, developed during a technical assessment, that predicts when employees might quit.