Mauricio Sánchez

Designing and implementing cloud and hybrid architectures to enable AI/ML solutions at scale. Working within the AI Architecture team, I build and manage infrastructure that supports document intelligence systems (OCR, IDP) running in on-premise environments and integrated with cloud services. I develop infrastructure as code using AWS CloudFormation and automate deployment workflows with Jenkins, enabling rapid and reliable environments for machine learning teams. My work focuses on enabling the development and deployment of LLM-based solutions, integrating multiple providers (AWS Bedrock, Azure OpenAI) and diverse data sources, ensuring scalability, security, and maintainability across hybrid cloud environments.

Designed and implemented complex data pipelines integrating data from databases, APIs, and external systems into the Databricks Lakehouse Platform. Built efficient data models for BI analytics and retail behavior programs using Python orchestrated by Azure Databricks Jobs and Snowflake Scripting. Developed robust data models and tailored solutions using Azure Data Factory. Collaborated with data scientists and business teams to improve data freshness, stakeholder trust, and business decision-making.

Implemented data quality controls, governance, and best practices across multiple data marts in BigQuery. Established monitoring and alerting systems using GCP services (Cloud Monitoring, Cloud Logging, Pub/Sub). Architected high-performance ETL pipelines using BigQuery Scripting, Airflow (Cloud Composer), and Apache NiFi. Designed the first DBT data models supporting the Medallion Architecture (Bronze, Silver, Gold layers). Performed root cause analyses in production, maintained documentation, and created architecture diagrams to strengthen data governance.

Created and maintained a data warehouse in PostgreSQL (Azure Database) for analytical and reporting needs. Developed and monitored ETL workflows using Airflow on Azure Container Instances with CI/CD via GitHub Actions. Automated Excel-based reports using Azure Functions, reducing manual processing time from two hours to under one minute. Supported deployment pipelines for production-grade ML APIs using Azure Container Instances and Azure Kubernetes Service.