Building AI products
with impact at scale
AI Product Manager. I build AI agents and the evaluation systems that prove they work.
Harvard MBA 2027, previously Lead Data Scientist at Colombia’s largest bank.
Design the system
I built the internal ML platform 153 teams use to put models into production.
It automated 600+ models. Standardizing the path was the easy half; getting teams to adopt it was the product problem.
Measure first
I designed an experimentation dashboard that cut exploration cycles for 3 priority projects in half.
Data Science and Business finally argued from the same numbers. Evidence beats opinion.
Prove it in production
I owned a real-time fraud model that scored 4M transactions a day and prevented $700K a month.
It had to run at transaction speed. Three data pipelines over 1.8B transactions grew the model's coverage 1,100%.
Building impactful products and sharing what I learn along the way
See the work. The decisions behind the numbers, and what I’d do differently.
About Me
I started as an engineer, became a data scientist, and now build AI products.
I spent four years at Bancolombia, Colombia’s largest bank, taking ML products from problem definition to production impact, ending as Lead Data Scientist. I’m now an MBA candidate at Harvard Business School, class of 2027.
My view: as models commoditize, measurement is the moat. Evaluations and metrics from day one are what make an AI product trustworthy.
More about me:
- I have practiced Calisthenics for 12 years. Front lever is my top skill.
- I’m a metalcore and live music fan. My favorite band is Imminence.
- I love Nintendo games. My current goal is to get all stars in Overcooked 2.