About
I build AI and data systems that turn raw data into decisions.
Carlos Ferrer Genao
Data Scientist | Data Engineer | AI & Analytics Developer
New York, NY
I'm a data scientist and generative AI specialist with 5+ years across AI/ML, analytics, payments and fraud. I hold an M.S. in Data Science from Fordham University and a B.S. in Software Engineering from INTEC, and I build AI and data systems that businesses actually run on.
As a Data Scientist at Auto Valley Group, I built an AI agent that uses OpenAI tool calling across 21 analytics tools to write weekly reports for non-technical stakeholders. Every number it produces is traced back to source data before anyone reads it. I also built the company's market intelligence platform (PostgreSQL, the MarketCheck API and Next.js), which syncs competitor data twice a day, and a prescriptive decision engine in production that turns market data into an explainable opportunity score, target buy and retail prices, and acquisition plans that stay within budget.
I learned data where accuracy matters. At Banco Popular Dominicano I investigated 200+ fraud cases a month across 1M+ monthly transactions in SQL, improving the detection rate by 25% and cutting false positives by 30%. As a Project Leader there, I delivered a new payment process for one of the largest supermarket chains, processing 10,000+ transactions a month with 40% faster reconciliation. At NeuronBrains, I built executive Power BI dashboards and automated the reporting behind them.
My toolkit spans the whole lifecycle: Python, SQL and PySpark for data work; PostgreSQL, ETL/ELT pipelines and data modeling for storage; XGBoost, CatBoost, SHAP and BERT for modeling; and ChatGPT Enterprise, Claude, the OpenAI and Anthropic APIs and n8n for AI agents and automated workflows. I'm also the person colleagues come to with complex AI and technical problems, and I coach teams on prompt engineering, model selection and AI best practices.
What ties it together is turning a business problem into a practical solution: follow it from the raw source, through the pipeline and the model, to the screen where someone makes a decision, and make sure the numbers hold up along the way.
Technical interests
What I keep coming back to
Data engineering
Warehouses, pipelines and automated refreshes — the layer that decides whether everything downstream can be trusted.
Explainable ML
Models that come with their reasons attached, so the people using them can challenge the output.
Causal thinking
Separating what correlates from what actually moves an outcome — the difference between a forecast and a pricing decision.
Grounded AI
LLM features where language models plan and explain, and deterministic code produces the numbers.