Gary Zavaleta

Analytics & Data Science Advisor

A data-driven technologist, and AI strategist with over 15 years of experience leading analytics, and applied data science across banking, travel, hospitality, and fintech industries. A recognized leader in the UAE, Gary has designed and led multi-million-dollar programs that blend cutting-edge technology with pragmatic business value.

View GitHub Download Resume

About

He serves as Senior Director of Analytics & Data Science at Global Hotel Alliance (GHA), where he spearheads enterprise-wide AI, and data-platform modernization. Beyond GHA, Gary is an advisory board member in AI and decentralized technology ventures. His current research and projects explore Agentic AI architectures and Quantum Artificial Intelligence, bridging today’s applied AI systems with the quantum-ready future. He is particularly focused on optimizing neural network training times using quantum algorithms, advancing the convergence of classical and quantum machine learning.

Featured Projects

Chat‑2‑DB

A GenAI chatbot that converts natural language to SQL using Retrieval‑Augmented and Table‑Augmented Generation methods. The project demonstrates how to democratise access to structured data, combining LLM providers with vector search.

Employs both RAG and TAG to generate SQL from natural language questions
Supports multiple LLM providers and vector search with PGVector for retrieval
Includes an evaluation framework and observability via Langfuse analytics

BigQuery RAG Text‑to‑SQL Gradio Python

Repo Resource

Ethereum RAG with BigQuery

An end‑to‑end retrieval‑augmented generation pipeline over Ethereum blockchain data stored in BigQuery. It demonstrates how to explore transaction and contract metadata using natural language and provides sample queries and data modelling best practices.

Queries Ethereum blockchain data in BigQuery using conversational prompts
Constructs a retrieval pipeline over transaction and contract metadata
Provides a Colab notebook with sample queries and an end‑to‑end demo

Ethereum BigQuery RAG Data Modelling

Repo Resource

Publications

Why 90% Accuracy in Text-to-SQL is 100% Useless

Designing and evaluating a text‑to‑SQL pipeline over BigQuery with retrieval and table augmentation guardrails.

Towards Data Science • 2026‑01‑12
Read more
Querying Ethereum on BigQuery: a Practical Guide

How to explore the Ethereum dataset, write queries, and build retrieval‑augmented metadata over blockchain tables.

GitHub Self Publish • 2025‑09‑22
Read more
Building Token Recommenders on GCP

Frameworks and best practices for building recommendation engines on Google Cloud Platform using BigQuery and AI Platform.

Blog • 2024‑03

Philanthropic Work

DataKind.org

08/2023 - Present

Volunteer Data Scientist for social inclusion projects

• Expanding Digital Financial Services in Afghanistan - Designed unsupervised models for customer segmentation, enabling improved targeting of financial services in underserved regions.
• Financial Inclusion in Colombia - Conducted exploratory analysis across multiple demographic and financial datasets and delivered insights to support evidence-based decision-making.
• Humanitarian AI Assistant - Conducted a technical assessment of DataKind's data-recipes-ai repository, evaluating architectural design, computational efficiency, scalability, and providing strategic recommendations for future development.

Get in Touch

I am always interested in collaborating on exciting analytics and AI projects, speaking at events or mentoring. Feel free to reach out via email or connect through social media.

Send Email Connect on LinkedIn