2026-02-12
Context
In the inventory planning and distribution workflows of Vinos América, there was a requirement to accurately automate the demand forecasting of thousands of combinations of product_id + warehouse_id.
To solve this challenge in a robust and industrialized way, VA_Proyecciones was designed and implemented, an internal modular Python library specializing in demand classification, statistical modeling, and backtesting of time series.
[!NOTE]
Confidentiality and Intellectual Property Note:
Being a predictive analytics library at the core of the company’s business, the source code and connection variables are not public (not open-source) in order to protect formulas and the organization’s competitive advantages in distribution.
2026-01-27
Context
The CineMX Yearbook of the Mexican Institute of Cinematography (IMCINE) constituted one of the most valuable public resources to audit, study, and research the state of cinema in Mexico. In late September 2025, I carried out an exhaustive data extraction (web scraping) process of the site in order to analyze thematic trends in Mexican films.
On January 27, 2026, the official online platform that allowed interactive query of these historical files was unexpectedly taken down. In order to prevent the loss of this historical information and encourage cultural and public data analysis, I decided to consolidate and open everything extracted through a public repository on GitHub.
2024-11-20
Context
Before the global development of the Model Context Protocol (MCP) ecosystem, there was a need to connect language models (specifically OpenAI’s Custom GPTs via Actions) quickly, securely, and directly with the company’s transactional system to enable real-time queries.
To solve this need, Vinos América Data API was designed and built, a robust REST interface that translates the AI’s natural language requests into secure, high-performance SQL queries against the SAP HANA enterprise database.