Director of Data Science • AI Product Leader

I build AI products that shorten the distance between effort and impact.

Most AI initiatives fail not because the technology doesn't work, but because teams optimize model performance instead of business outcomes. I fix that.

See my work ↓ LinkedIn Get in touch
10+
Years in Tech
$190K
Annual AI Impact
$350K
Engineering Savings
4
Countries Worked In

About

From engineering trenches to AI product strategy

I started as an engineering intern pulling cables at Siemens in Bogotá. A decade later, I lead AI teams at the intersection of data science, product, and business impact—having built and shipped across fintech (Nubank, Mercado Libre), logistics (Mensajeros Urbanos), and martech (Power Digital Marketing).

“Your job isn't to maximize intelligence. It's to shorten the distance between effort and impact.”

My career arc traces a deliberate evolution: traditional engineering → operations & analytics → data science leadership → AI product strategy. At each stage, I've carried one conviction: technology only matters when it changes decisions, behaviors, or results.

I hold a Master's in Data Science from Universidad Austral (Buenos Aires) and specialized AI training from MIT. I practice mindfulness, play golf since I was five, and believe the best teams are built on trust, clarity, and an unflinching bias toward action.

Impact

Proof of Value, not Proof of Concept

I don't believe in demos that don't pay invoices. Here's what actually shipped and moved the needle.

$190K/yr
AI Slackbot — Enterprise Knowledge Automation
Built an AI-powered Q&A bot across 9 channels, answering 528+ questions/month. Proved value with 80% accuracy in 4 weeks, optimized after adoption—not before.
Power Digital Marketing
$600K
Google Partnership — Meridian MMM + Vision AI
Led operationalization of Google's Meridian Media Mix Modeling with event-driven, GPU-backed architecture on AWS. Secured $38K partnership milestone.
Power Digital × Google
$350K
Contract & Design Optimization
Negotiated $150K in contract savings and delivered $200K in engineering design optimization as Lead Engineer for the Argentina-Uruguay region.
Siemens
+8% / -30%
Operational KPI Transformation
Improved OnTime delivery by 8%, reduced Unity Cost by 30%, and decreased Cancel Rate by 15% through data-driven operations and OKR implementation.
Mensajeros Urbanos
ETA Model
Last Mile Logistics Prediction
End-to-end ML product for delivery time estimation, generalizing across diverse scenarios using order characteristics, historical data, and real-time operations.
Mensajeros Urbanos
Share of Wallet
Seller Revenue Intelligence
Built marketing science models calculating monthly size, share, and potential of sales for active sellers using in-person payments as proxy estimators.
Mercado Libre

How I Think

Mental models that keep AI grounded

Frameworks I apply daily to cut through noise and increase the odds of real impact.

⌖
First Principles
Strip away the UI, prompts, and novelty. What's left must still change a decision, behavior, or result.
⇒
Effort → Impact Chain
High effort with low impact is failure. Low effort with high impact is excellence. Stop celebrating effort.
&diamf;
Proof of Value
Don't ask "Is this feasible?" Ask "If this works, does it matter?" Kill ideas early based on lack of value.

Journey

A decade of building across industries and continents

2024 – Present
Director of Data Science
Power Digital Marketing • Remote
Leading AI/ML team of 6+. Built AI Slackbot ($190K impact), operationalized Meridian MMM, secured $600K Google partnership. Driving AI productization aligned to P&L impact.
AI Products Team of 6+
2023 – 2024
Project Lead, Data Science
Mercado Libre • Bogotá
Led "Share of Wallet" marketing science initiative—calculating revenue potential for active sellers across LATAM's largest e-commerce platform.
Marketing Science
2022 – 2023
Data Science Manager
Nubank • Bogotá
Managed data science team at Latin America's largest digital bank, applying ML to fintech at scale.
Fintech People Management
2021 – 2022
Director of AI → Regional Ops Manager
Mensajeros Urbanos • Bogotá
Built ETA prediction models for last-mile logistics. Improved OnTime +8%, reduced costs -30%, slashed cancellations -15%. Implemented OKR framework company-wide.
Operations Research MLOps
2015 – 2021
Intern → Engineer → Sr. Applications Engineer
Siemens • Colombia & Argentina
6-year journey from intern to Lead Engineer for Argentina-Uruguay. Delivered $350K in savings. Won Innovation Tank LA Award 2020.
$350K Savings Innovation Award

Education

MIT + Universidad Austral

MIT Professional Education
Data Science & ML, AI for Business, Computational Thinking
2021 – 2022
Universidad Austral
Master's in Data Science
2019 – 2020 • Buenos Aires

Writing

Thinking in public

I write about the gap between AI model accuracy and real business impact, measuring data science teams, and building products that people actually use.

Building AI Products That Actually Matter
Mental models for shortening the distance between effort and impact
→
Data Science Dynamics: Balancing Business Impact and Team Productivity
Framework for measuring DS team productivity aligned with business goals
→
Empowering Data Science Teams: Strategies for Maximum Impact
Measuring DS team performance by impact, not output
→
Interacting with ChatGPT: Exploring Prompting Techniques
Zero-shot, multi-shot, role-acting, and knowledge integration
→

Connect

Let's build something that matters

I'm always interested in conversations about AI product strategy, data science leadership, and closing the gap between what we build and what gets used.

jaime.diazbeltran@gmail.com LinkedIn Medium