Kueski
Senior Data Scientist
Owned production fraud ML end to end. Cut false positives by 65% and fraud losses by 70%, and added graph analysis to surface fraud rings that single-account models kept missing.
Vancouver, BC
Data Science / AI / Products / Entrepreneurship
I turn ideas into products.
I've spent the last several years moving between startups, machine learning systems, growth experiments, AI products, and a few businesses of my own. What ties it together is figuring out how things work, how they grow, and where data and AI can have a meaningful impact.
What keeps me interested is the possibility of turning ideas into things that people actually use. Whether it's a fraud model evaluating thousands of decisions every day or a product reaching its first users, I find the greatest satisfaction in seeing technology move beyond prototypes and into the real world.
01Featured product
My own product
Live product · Founder
brewo
An AI-powered newsletter for Latin American startups, tech & finance
Brewo began with a simple question: what if a small crew of AI agents could research and write a genuinely good newsletter every day? So I built it. The idea, the product, and the pipeline behind each issue, covering startups, tech, and finance across Latin America.
- A small crew of AI agents that research, summarize, translate, and draft each issue.
- The whole product around it: site, subscriptions, payments, and delivery.
- Constant tinkering based on what subscribers actually open and read.
brewo
AI-brewed startups, tech & finance for LATAM
Gen AIRAGMulti-agent workflowsProduct developmentGrowth
02
Projects
A mix of things I've built, some at work and some on my own. They look unrelated until you notice they're all me poking at the same question from different angles.
Founder
MunchGo, a multi-brand ghost kitchen
A delivery-only kitchen in Mexico City running three brands out of one space: MunchGo, WingsEat, and Munchin' Dogs. I chose the menus, pricing, and positioning by digging through Yelp and maps data to find gaps the competition had left open.
03
Where I've worked
Startups, a unicorn, and a global brand. Different settings, but mostly the same job: find where data and AI can help a business grow.
Kueski
Senior Data Scientist
Rappi
Growth Data Scientist
Heineken
Growth Data Scientist
Insaite
Data Science Sales Engineer
Axity
Junior Data Scientist
04
About me
Hi, I'm Elí.
I'm a data scientist, AI engineer, and entrepreneur who enjoys turning messy questions into useful products, systems, and decisions.
Over the years I've worked across fraud prevention, growth experimentation, machine learning, and AI products in startups, fintech, marketplaces, and consumer businesses. What keeps me interested is the challenge of understanding a problem, identifying what matters, and turning ideas into something people actually use.
The work I'm most proud of isn't necessarily the most technical. It's the products, models, and systems that ended up influencing real decisions and reaching real people, whether that's a fraud model evaluating thousands of applications, an experiment shaping product strategy, or an AI product finding its first users.
When I'm not working, I'm usually building something: AI products, open-source software, side businesses, or writing about what I learn along the way.
// profile
05
Work
Six years across fintech, a Latin American unicorn, and a global consumer brand. A few of the problems I got to chew on along the way.
Heineken
Growth Data Scientist
Ran a subscription experiment that lifted revenue 16%, and built an NLP pipeline that sorted incoming customer reviews to the right team in under two minutes.
Rappi
Growth Data Scientist
Built marketing attribution to see which channels actually paid off, and ran A/B tests on vendors and creatives with causal impact analysis to separate signal from noise.
Insaite
Data Science Sales Engineer
Sat between data science and the sales conversation, scoping ML solutions and explaining, in plain terms, what they'd actually do for a C-level client's business.
Axity
Junior Data Scientist
My first role in data. EDA and feature engineering on client projects, where I learned to ask the right questions before reaching for a model.
06
Education
Master of Data Science
University of British Columbia · Vancouver, BC · 2025–2026
Data Science & Machine Learning Certification
Massachusetts Institute of Technology · 2022
Bachelor of Actuarial Sciences
Universidad Nacional Autónoma de México · Mexico City · 2020
07
What I work on
Less a list of tools, more the handful of areas I keep coming back to, and how they tend to feed into each other.
Machine Learning
Training models that hold up outside a notebook: fraud, risk, NLP, and the unglamorous production parts that decide whether any of it matters.
Pythonscikit-learnFraud & riskNLPFeature engineering
Generative AI
LLM-powered products and agentic workflows. Fun precisely because the whole field is still half figured out.
LLMsRAGMulti-agent workflowsClaude / GPTEvals
Experimentation & Causal Inference
A/B tests, causal impact, and the constant, healthy fight against fooling yourself with a misleading chart.
A/B testingCausal inferenceAttributionStatistics
Product Analytics
Understanding how a product is actually used, then turning that into decisions someone can act on.
SQLGrowth metricsDashboardsPlotly Dash
Product Development
Taking an idea all the way to something people can open in a browser and use.
Next.jsSupabaseStripeFull-stack
Entrepreneurship
Starting things, pricing them, and learning the parts of a business the spreadsheet never warns you about.
FoundingOperationsGo-to-marketPricing
Open Source Software
Packaging the useful bits of my work so other people can pick them up and skip the boring part.
Python packagingPyPIDocsTesting
08
Right now
Doing my Master of Data Science at UBC, digging into ML systems, LLM applications, and the analytics that help products and teams make better calls. Still building things on the side.