Money lost on every route to spoilage and theft, until we started tracking every truck live.
Open water swimming taught me to love hard things.

02The work · Bogotá, 2025–2026
Then I learned to sell what you can't see.
Oberon TSI is a risk consulting firm in Colombia: companies hire it to find and fix the leaks in their operations. I led B2B product and brand marketing with a team of three designers. The engineers and analysts built the technology; my role sat between their analytics and each client's strategy, turning it into products clients could understand, use and buy.

What the work moved.
We tracked all the hospital waiting rooms live and found the leak, so they knew what to fix first.
Clients couldn't tell 20+ services apart. I grouped them into 3 brands and improved SEO.
Clients didn't know these services existed. I turned them into LUNA and Oberon News.
See the risk before it costs money.
Oberon 360 is an app that puts cameras, GPS and sensors on one live screen, so clients see real-time data on their products, fleets, logistics and operations. I positioned it and brought it to clients.
Client names and personal data are blurred.
LUNA
Warns clients about risks before they happen. I gave it a name, a face and a launch.

Oberon News
Weekly risk alerts that give clients a heads-up before something goes wrong. I named it, branded it and launched it with the team.




My first marketing job came from sports.
For a year at a sports-hydration brand, I managed athlete ambassadors and team sponsorships in triathlon, running, swimming, cycling, BMX, tennis and soccer. I also created content and ran the brand at races where it was the official hydration partner.
I was in charge of
And I built a race for the kids who couldn't run it.
For a nonprofit organization, I created Colombia's first charity race for children with cancer: Abraza un sueño, corre con el corazón. Kids from the foundation hung the medals on the runners, and those who had finished treatment rang the bell that marks the end of it.





Intuition is a hypothesis.
Data decides whether it ships.
I've been the most confident person in the room, and I've been wrong. Working closely with analytics and engineering teams taught me to test instead, so every idea goes through four steps.
- 01Pick the metric before the campaign.
- 02Run it.
- 03Read the result honestly.
- 04Reallocate.
05Now · McCombs School of Business
Now I'm learning to predict, not just report.
The more I worked with data, the more I wanted the tools to analyze it myself. So I came to Texas McCombs for my Master in Marketing Analytics. Here I'm learning to model what customers prefer, find the segments that matter, predict what they'll do next and measure what marketing really moves.
- Data analysisR, Python, SQL
- Statistical modelingLinear and logistic regression, ANOVA, A/B testing, hierarchical models
- Segmentation & choiceCluster and factor analysis, conjoint, discrete choice models
- Customer & pricing analyticsCLV, RFM, price elasticity, pricing optimization
- Data visualizationTableau, Power BI, ggplot2
- AI & automationClaude Code, Codex

- Marketing analytics
- Product & consumer insights
- Market research
- Customer analytics & CRM
- Performance marketing
- Demand generation
- Growth
And if it's in sports, even better.


What I’m building at McCombs.
Projects from the master’s, added as they happen. Each one started as a question in class and ended as something people can use: research, tools, and whatever comes next.



The brand gave us no data. We built our own.
Case partner: Lucchese Bootmaker, a Texas Western-boot brand with more than a century of history. The question: how does a premium heritage brand grow beyond its core customer without discounting? With no customer data, our team of four designed and fielded two original surveys (n = 162) and built the recommendation on them. My role was research design and analytics: I wrote the survey instruments and ran the analyses in R: a choice-based conjoint with willingness-to-pay, a driver regression, factor analysis and k-means segmentation, a randomized message A/B test and a TURF reach analysis. Together we turned the findings into the strategy, the financials and the story we told the judges.
Methods only: the brand’s results, the survey data and the deck stay with the competition.

An agent that reads the job market for me.
Every weekday morning a scheduled agent scans marketing openings in the United States and Colombia, maps each one to the eight areas of marketing and the McCombs curriculum, and scores it 0–100 against my resume and salary targets. I built it with Claude Code, Claude Projects and a set of skills, and shipped the public version as a single HTML file: the Marketing Job Search Finder. A paid edition, the AI-Powered Career Navigator, lets anyone match their own program and resume to real postings.

