At UNICRED, I’m the technical owner of a unified data, BI and MLOps platform on Azure Databricks. It supports regulatory credit-risk models and around 1,000 orchestrated tasks serving Central Bank reporting, BI and ML.
I brought the regulatory provisioning simulation from 50 hours to 10 minutes, and it’s now in production. The platform versions 36 model revisions and is guarded by around 1,550 automated tests.
I built an LLMOps capability with specialized agents, guardrails and an evaluation-driven improvement loop. Nine AI agents use seven domain MCP servers for typed, bounded access to the data catalog, regulatory rules, job runs and data.
Before UNICRED, at Cooperativa Central Ailos, I developed and maintained models on Databricks and Spark, including work with datasets containing more than 300,000 features, and migrated legacy SAS analytics to Spark. Earlier, I worked in surveying and geomatics engineering; I’m a Databricks Certified Data Engineer Professional.

