Now shipping · Available Summer 2027
Fabio Macedo

Fabio Macedo · MIT Sloan MBA ’28 · 6 years at AB InBev, 1 as Product Manager

Building platforms that pay off for both sides.

At BEES, AB InBev’s B2B commerce app for 300,000 shopkeepers in Peru, I fixed what the product recommendation model was optimizing for (+17.1% first-time purchases), built the first monetization strategy ($2.6M EBITDA in year one), and created a co-funding model that adds $100 a month to 10,000+ shopkeepers’ income.

Builds with AI~20 LLM-assisted A/B tests at BEES, and a Claude Code bot at Sloan that 5 classmates now run.

Email me LinkedIn CV +1 (857) 250-8074Call
MIT Sloan MBA ’28 Sloan PM Club VP of Recruiting Organizing Sloan Tech Summit 2027 Dean’s Fellow Former PM at AB InBev
Where I’ve worked and studied
  • AB InBevBEES
  • The Adecco Group
  • MIT SloanMBA ’28
  • Universidad del PacíficoBA

By the numbers · impact

Six years, four roles, numbers that moved.

$2.2M

Promotional budget moved to volume-driving discounts after ~20 LLM-assisted A/B tests

PM, Online Strategy

8M+/mo

Recommendations shown by the ML model I re-pointed at harder sales

PM, Online Strategy

+17.1%

First-time purchases, H1 2026 vs. H1 2025, after the new model launched

Case study ↓

10K+

Shopkeepers on a co-funding model I created with no precedent, adding $100/mo to their income

Monetization Manager

$2.6M

Year-one EBITDA from BEES Peru’s first monetization strategy, 13 partner companies

Digitalization Lead

100%

Promotion rate across my three direct reports in 2024

People leadership

What I bring on day one

What I bring
How I work
Where I’ve shown it
Trade-offs, made on purpose
How I workI use data to decide what to optimize, and I’m explicit about what it costs.
Where I’ve shown itCoverage vs. precision across 8M+ monthly recommendations, settled with Data Science and shipped with Engineering.
A view of the whole business
How I workI read a product through its P&L: pricing, loyalty and monetization.
Where I’ve shown itA new monetization model that built $2.6M in year-one EBITDA and strengthened loyalty at the same time.
One shared vision of success
How I workI bring people on board and align them on what success looks like.
Where I’ve shown itSkeptical bosses won over with data, C‑suite buy-in for the loyalty redesign, and 1,000+ sales reps trained to use what we shipped.
Customer obsession
How I workI start with the customer: what they value, what problems they have, and what we can do about them.
Where I’ve shown itA co-funding model built on shopkeepers’ unmet needs, adding $100 a month to their income.
A good person to work with
How I workI put people first. Empathy and shared goals are how a team grows together.
Where I’ve shown itIn 2024, all three of my direct reports were promoted. Managers, peers and reports describe me the same way: empathetic, collaborative, someone who lifts the team.

Case study · ML recommendations

The model was rewarding the easy sale. I made it chase the hard one.

BEES personalized offers for ~50 SKUs across 300,000 shopkeepers, at the SKU × customer × lever level (direct discounts and loyalty points). By 2025 the recommendation model had learned the wrong lesson, and our sales incentives were making it worse.

  1. The hypothesis that failed

    My first read: we were losing distribution on the SKUs shopkeepers always buy. The data said no. Those were our best performers. I dropped the idea and kept digging.

  2. The real problem: incentives

    The model spent its biggest discounts on recommendations that would have converted anyway. Sales reps were paid on distribution (avg. SKUs per customer × customers), and they could hit it by converting only the easy recommendations.

    Reps hit their variable pay. The company missed its goals. Medium and hard recommendations almost never converted.

  3. The call: coverage over precision

    I proposed flipping the model: put the aggressive discounts and top-of-carousel visibility on harder-to-convert recommendations, and trust the easy ones to convert on their own.

    My boss disagreed at first. I showed him the data: reps were hitting variable pay while the business was missing its goals. He signed off.

  4. Shipping it with Data Science & Engineering

    I led the iteration. Data Science and Engineering built the prototype; I pressure-tested the output and drove two more working sessions of changes. v1 went live January 1. We kept fixing smaller issues until the final model shipped at the end of February.

  5. What moved

    +17.1%First-time purchases, H1 2026 vs. H1 2025
    +7.6%Distinct SKUs sold, year over year
    3% → 10%Medium/hard recommendations’ share of distinct SKUs sold
    #6+ → #1–2Carousel position of medium/hard recommendations

How we measured it All results are year over year (H1 2026 vs. H1 2025). The new model was the only change to distribution that year, and the 10% of shopkeepers we kept on the old model as a control group grew close to zero.

What I took from it The model was doing exactly what we rewarded. When a metric is hit and the goal is missed, look at the incentives before the algorithm.
Try it · BEES home carousel
Recommended for you5 slots, then “see more”
See more → Behind “see more”
Easy−22%−6%
Easy−20%−5%
Easy−18%−5%
Easy−18%−4%
Easy−15%−4%
Hard−4%−25%
Med−5%−20%
Hard−3%−22%
Med−4%−18%
−20% discount offered on that product Easy · Med · Hard how likely the shopkeeper was to buy it anyway
3%10%

Medium/hard share of distinct SKUs sold

#6+#1–2

Where medium/hard recommendations start in the carousel

Where the discount budget goesIllustrative
Easy to convert Medium Hard

Real: medium/hard recommendations moving from #6+ to the top of the carousel, their share of SKUs sold, and the results. Illustrative: the specific products, discounts and budget split, which varied by SKU and customer.

Product decision · loyalty rewards

Rewards nobody saw. I moved them to where people buy.

The problem. Loyalty rewards lived in one section of the app that shopkeepers didn’t need to visit to place an order, and they had to “accept” a promotion to earn the points. Most never saw them.

The decision. Surface the rewards in the steps every order already goes through: the home page, the product page and checkout. I defined the user stories and requirements with Analytics and Design.

44% → 83%Share of users earning rewards, 2022 → 2023. No control group, but it was the only major loyalty release that year.

Reviews · managers and reports

Reviewed by my managers and my team.

from a recommendation letter
I’ve had the opportunity to lead many high-performing individuals, but Fabio stands out as an exceptionally well-rounded profile.

“He gets deeply involved, asks relevant questions that bring blind spots to light, and frictionlessly challenges ongoing conventions.”

“He has my utmost trust; not only is he my go-to proxy whenever I’m out of the office, but he also stepped up and led my team flawlessly for 8 months in 2024.”

Mauricio SialerOnline Strategy Senior Manager, AB InBev
Manager
via LinkedInManager
“He has an exceptional ability to break down complex challenges, identify priorities, and implement effective solutions that have directly impacted not just our area’s results but the company’s overall performance.”
Marcela Vera SolfEx-AB InBev & BEES · PMP
via LinkedInReport
“He leads a team with clarity and motivation… ensuring that each team member feels valued and supported.”
Rodrigo Begazo YañezKey Account Supervisor

Quoted from a recommendation letter and public LinkedIn recommendations. Read all on LinkedIn ↗

Changelog · experience

Six years at AB InBev, the last as Product Manager.

I joined AB InBev as a Global Management Trainee and left as Product Manager for BEES Peru’s online strategy. Every release shipped with numbers.

  1. v5.0
    2026 – present
    In development

    MBA Candidate (STEM), Dean’s Fellow

    MIT Sloan School of Management · Cambridge, MA · May 2028

    LeadershipProduct Management Club

    VP of Recruiting

    I help Sloan students navigate recruiting for product roles: how companies open positions, when to apply, and how to get ready for interviews.

    • Info sessions
    • Company visits: Google, Toast
    • Peer interview groups
    • 1:1 interview coaching
    • Guest speakers
    • AddedSloan Tech Summit 2027, sponsorships committee: working with the content team to match companies with what the conference offers (networking booths, panel seats, company talks, resume books).
    • AddedMember of the LATAM Club and Sloan FC.
    • AddedCoursework in Product Management Lab, AI Builder Seminar and Communicating with Data.
    • AddedCertificates in Product Management and Business Analytics (in progress). Building with AI on the side; see Labs.
    Fabio taking a selfie in front of the MIT Sloan School of Management sign
    Week one. No regrets.
  2. v4.0
    2025 – 2026

    Product Manager, Online Strategy

    BEES · AB InBev · Lima, Peru · B2B e-commerce marketplace serving 300K+ shopkeepers

    • ShippedML product recommendation model with Data Science and Engineering, balancing coverage vs. precision across 8M+ recommendations a month. +17.1% first-time purchases and +7.6% distinct SKUs sold, year over year. Case study ↑
    • AddedAn LLM-assisted A/B testing workflow in Microsoft Copilot. From a test template (variables to hold fixed and to test), it designed and split several test groups against one control, then read out uplift and statistical significance; Data Science validated every readout. About 5 tests a month for 4 months comparing discounts with loyalty points moved $2.2M into volume-driving promotions, in a year company volume grew 4%.
    • ImprovedRedesigned the loyalty program’s reward-allocation logic and secured C-suite buy-in. Share of investment driving incremental revenue: 74% → 85% YoY.
    • ImprovedBrought competitive-strategy data into recommendations for the first time, with Data Science, Revenue Management and Sales, contributing to a +1 pp market share gain (H1 2026 vs. H1 2025).
  3. v3.0
    2024 – 2025

    Monetization Manager

    BEES · AB InBev · Lima, Peru

    • AddedA co-funding model for shopkeepers with no precedent, built from research into unmet needs. Scaled to 10K+ shopkeepers and +$100/mo in their income, helping monetization reach an all-time high of $3M in 2024 (+15% YoY).
    • ShippedPay with Points: shopkeepers pay for orders with loyalty points at a better conversion rate. Monetization +18% in 2025.
    • AddedBEES School: bi-weekly training for 1,000+ sales reps on new features and adoption. Reps’ adherence to recommended app use +15 pp.
    • ImprovedManaged three direct reports on personalized growth plans: 100% promotion rate in 2024. Led my manager’s team as acting lead for 8 months in 2024.
  4. v2.0
    2021 – 2023

    BEES Digitalization Lead

    BEES · AB InBev · Lima, Peru

    • ShippedDefined user stories and requirements with Analytics and Design for a feature that made loyalty rewards more visible. Users earning rewards: 44% → 83% (2022 → 2023).
    • AddedLed BEES Peru’s first third-party distribution deal end to end, from first contact through negotiation, pricing and launch. $500K GMV in year one.
    • AddedBEES Peru’s first monetization strategy: positioning, partner messaging and pricing built on partner and customer needs. 13 companies onboarded, $2.6M EBITDA in year one.
  5. v1.0
    2020 – 2021

    Global Management Trainee

    AB InBev · the company’s top leadership development program

    • AddedSelected as 1 of 3 from 1,700+ applicants; rotations across marketing, sales and supply chain.
    • ShippedDrove the creative process behind an in-house digital content creator for the beer category. 5M+ views, 1M+ likes, 200K+ followers on Instagram and TikTok in three months.
  6. v0.9
    2019 – 2020

    Continuous Improvement Analyst

    The Adecco Group · Lima, Peru

    • ShippedCaptained a 17-person team using design thinking. Four process redesigns adopted company-wide.
    • ImprovedCut waiting times by 20% for 500+ daily interview candidates by redesigning registration and queuing.
  7. Edu
    2015 – 2020

    BA in Business

    Universidad del Pacífico · Lima, Peru · exchange at King’s College London

    • AddedTop 10 of a 139-person class.
    • ShippedFirst PM role: product management intern at Belcorp, helping build a sales-force training app.
    • ShippedLed a pro bono consulting team for a mental-health NGO, doubling yearly donations by rebuilding how they collected and analyzed data.

Roadmap · objectives

Built for 300,000 shopkeepers. Next: millions of people.

I want to be the PM who always delivers the numbers, but whom people remember for making their lives better.

At BEES, I worked to improve the income of 300,000 families who own small shops in Peru. It’s why I designed Club B Black’s benefits (legal counsel, telemedicine and e-learning) for shopkeepers and their families, not just their businesses.

Now I want that work at a larger scale: products that create value for the company and for millions of people. That could be saving them time, giving them a better experience, or, like at BEES, helping them earn more to provide for their families. Product is the best way I know to do that.

ProfessionalLong term

Products used by millions

Take what worked for 300,000 shopkeepers to products millions of people use: saving them time, giving them a better experience, helping them earn more.

Results people remember

Keep delivering the numbers, and make sure the work improves someone’s life along the way.

Build marketplaces and commerce at scale

Two-sided platforms where getting incentives right moves both the business and the people on it. Gaming, film and TV are the products I love as a user.

LearningIn progress

MIT Sloan MBA, Class of 2028

Product Management Lab, AI Builder Seminar, and the Product Management and Business Analytics certificates.

Building with AI

Shipping small tools for my own workflow, to keep the build half of the job sharp. See Labs.

PersonalOngoing

Grow the people around me

Teams where everyone develops, like the three reports who were all promoted in 2024 and the 1,000+ reps trained through BEES School.

Labs · side projects

I build the tools I wish I had.

Small, real products for my own life at Sloan, built with AI. They’re where I practice the build half of the job.

Shipped

Sloan task simplifier

Problem: deadlines are spread across Canvas, clubs and recruiting, and it’s easy to miss one.

Build: a bot I made with Claude Code that runs on my laptop, scans Canvas every day and emails me what’s due in the coming week, from Gmail to my MIT inbox.

5 classmates run it today; 15+ asked for the build guide.

View on GitHub
Next build

Sloan “brain”

One place for two years of notes, cases and contacts, so I can come back years later and find what I need.

Decision: an LLM I can ask, not a repository I browse. Years from now I’ll remember the question (“what did we learn about pricing?”), not which folder the answer is in.

Live

This website

I directed it like a product: goals, content, design calls and several feedback rounds, including a recruiter review. Claude Code wrote the code; GitHub publishes every change automatically.

How I built it

Interests · outside of work

What I do when I’m not shipping.

Arsenal FC, #1 fan

It started with Winning Eleven and PES. I always picked Arsenal because it was first in alphabetical order. After a few years of playing as them, I started watching the real games, and then understanding what the club is about: achieving things the right way. Along the way I made my dad a fan too, and we’ve watched them together at the Emirates.

Victoria Concordia CrescitVictory through harmony · the club’s motto, and now mine

In an Arsenal scarf high in the stands before my first Arsenal game, against Watford
First Arsenal game! 2-0 win vs. Watford · 2018
A selfie with my dad in the stands at the Emirates, both in Arsenal shirts
Back at the Emirates with my dad! 3-1 vs. Nottingham Forest · 2024
Celebrating with my dad at home in Lima as Arsenal lift the Community Shield on TV
Community Shield winners!! My last day living in Lima · 2026

Amateur chef, two cuisines

Peruvian and Italian. If you want something good, ask me to cook: lomo saltado, always with a good pisco sour, or my bell pepper pasta.

Call of Duty: Zombies

How I stay in touch with my high school friends. Favorite map: the OG Kino der Toten, where my record is round 39. Call of the Dead is growing on me, though.

Future certified life coach

A dream for after the MBA: the ICF coaching certification. I wanted to be a psychologist before I chose business, and I’m still curious about that world.

Harvard WorldMUN 2018 champions

Head delegate of the Peruvian team that won Harvard World Model United Nations 2018. I also trained 80+ students for Model UN.

Lima → Cambridge

Native Spanish, professional English. I can read a room in two languages and catch what gets lost between them.

How I built this · tools & process

Built with Claude Code. Shipped by GitHub.

No template and no website builder. I treated my own site like a product: interview the user (me), write the spec, build it, ship it, and keep iterating.

Live sitefabio.cool
Source codegithub.com/fabiomn97/Who-I-Am
MeSet the goal, decided what to say and what to leave out, chose every story, number and photo, and directed the design through many rounds of feedback, including from a recruiter review.
Claude CodeInterviewed me, drafted the copy, and wrote the HTML and CSS. It also checked every version on phone and desktop before shipping.
GitHubHosts the code. A GitHub Actions workflow publishes it to GitHub Pages on every push, so an update goes live in about a minute.
~/who-i-am
❯ claude "Help me build a profile that
  showcases my history in an organized,
  intuitive way. Interview me first."

● Interview · what it asked me
  ├ What do you want to do next, and why?
  ├ Which story best shows how you work?
  ├ What was your part, and the team’s?
  ├ What failed, and what did you change?
  ├ Who can speak to your work?
  └ How should the page look and feel?

● Build
  ├ index.html    HTML + CSS, no framework
  └ checks        light/dark, mobile, a11y

❯ git push origin main
✓ GitHub Actions · deploy to Pages
✓ live at fabio.cool

Request a chat

Building something that matters? Let’s talk.

I’m looking for product roles where the hard part is real: messy users, genuine tradeoffs, numbers that have to move. I answer everything, recruiters very much included.

fabiom97@mit.eduEmail /in/fabiomacedonatersLinkedIn Fabio-Macedo-CV.pdfCV +1 (857) 250-8074Phone