Peru
I genuinely enjoy working with data, AI, and agents. They are the closest thing we have to Hari Seldon’s psychohistory, an attempt to decode reality at its deepest level and use that understanding to steer the wheel. That being said, my real motivation is helping institutions use these tools to evolve. I care about turning intuition into structured bets, assumptions into testable hypotheses, strategy into an intellectual craft. About redesigning and optimizing operating models so human effort can move away from repetition and toward creativity, innovation, and the arts. And then, once everything has been optimized, break things again, so we can start the cycle all over, reaching a new level of what is possible. I am continuously learning how to make this real within corporate environments, with the long-term goal of eventually applying those lessons to public policy. Because technology does not change society by itself, but rather by the institutions that learn how to use it better. I write occasional thoughts on Substack: https://substack.com/@alejandropalomino Share experiments on Github: https://github.com/AlejandroPSJ2 And some of my own music compositions on Youtube: https://www.youtube.com/@AlejandroPSJ
Led an interdisciplinary team (statisticians, engineers, economists) of data scientists dedicated to develop advanced analytics solutions with the use of big data. These involve the development, piloting (through randomized control trials or A/B testing), deployment and automation of machine learning algorithms. The predictive models cover several strategic areas for the company, as: client churn, receipt payment, marketing campaigns, people analytics (hiring), fraud prevention and detection, customer lifetime value, risk-based pricing, image/audio recognition, chatbots, etc. The development of these models involves an advanced use of Python, SQL, Power BI and Amazon Web Services and mastery of supervised (boosted trees, random forests, logistic regression, OLS, deep learning for classification) and unsupervised (clustering, look-alike, recommendation systems) methods.
Contributed to the elaboration of evaluation reports of IADB’s institutional / country / sectoral development strategies. Among the main reports elaborated were the sector notes of infrastructure, science and technology, trade and productive development for the Uruguay, Argentina, Perú and Guyana country-reports, where the IADB's portfolio was assessed. Developed algorithms to automate the processing of the IADB’s big data for the analysis of the 500 impact evaluation studies financed by the Bank in the last decade. These evaluations involved both experimental and quasi-experimental methods (difference in differences, propensity score matching, instrumental variables, regression discontinuity design, etc.). Contributed to the evaluation of the relevance, effectiveness and sustainability of these evaluations and its recommendations.