Paulo Brandao
Résumé

Platform & web product · APIs, SDKs, live ops

Paulo Brandao

The fastest way I know to win a product argument is to build the thing and put it on the table.

Eighteen years of platform and games product — ad tech SDKs, no-code creation tools, live ops. I write the specs and the roadmap like any PM. Then I go write the prototype, generate the art, and ship it.

$300M
AdColony exit, on a format I created
500%
DSP revenue growth at InMobi
150M+
Downloads across shipped titles
How I work What I've built

How I work

I've spent most of my career sitting between engineers who wanted a spec and executives who wanted a demo. This loop is what I settled on.

It used to take a quarter and three teams. With coding agents doing the typing, it takes about a week and one person — which changes what's worth arguing about in a roadmap meeting, and what's just worth building.

  1. 01

    Frame the problem

    Who's blocked, what it costs, and the number that will tell us it worked — written down before anyone builds.

  2. 02

    Build the argument

    Coding agents, a no-build stack, generated art. A working thing beats a deck in every review I've run.

  3. 03

    Hand it over clean

    Specs, flows and acceptance criteria engineering can price without a translation meeting. Sprints in Jira.

  4. 04

    Then argue with data

    Analytics and tests wired in at launch, so the next roadmap follows behavior instead of the loudest voice.

  • Product Strategy
  • Roadmapping
  • 0→1 Product
  • PRDs & Specs
  • Prioritization
  • User Research
  • A/B Testing
  • Analytics & Instrumentation
  • SDK & API Products
  • Developer Experience
  • Live Ops
  • F2P Economy Design
  • Agentic Development
  • AI Tooling & Prompting
  • Agile / Scrum
  • Go-to-Market
Code
JavaScript, HTML5/CSS, PHP 8, Node, MySQL/noSQL, Alpine, htmx, Tailwind, PixiJS, Three.js, Unity
AI
Claude Code agents, skills and MCPs · Claude, ChatGPT, Gemini · Zed, Antigravity, Replit · image generation
Data
Google Analytics, Mixpanel, Pendo, GameAnalytics, AppsFlyer
Ship
AWS, DigitalOcean, Railway, VPS · GitHub, Jira, Confluence, Notion, Figma

On AI

I was shipping generated content into production before it had a hype cycle attached.

Automated creative in an ad platform, then AI research for a game creation toolchain, and now agents doing the typing on everything I build. What I've learned in that stretch is less about the models than about where they move a product team's constraints — and which ones they don't move at all.

  1. 2020–22 · InMobi

    Generation inside the product

    Shipped Expresso, a creative generator built into the DSP so campaigns stopped waiting on a design queue. Same lesson as today: automate the bottleneck, not the interesting part.

  2. 2023–24 · Buildbox

    AI in a creation toolchain

    Came back as a consultant to launch Buildbox 4 and research what AI could realistically do for no-code game creation — separating the demos that impress from the features people would actually use.

  3. 2025 → · Now

    Agents doing the typing

    Claude Code agents with custom skills and MCP servers, generated art run through my own sprite pipeline, and research synthesis that used to be a week of reading. Everything on this page came out of that setup.

What it changes for a product team

The cost of being wrong drops.

When a prototype takes days instead of a quarter, you can test three directions instead of defending one.

One person covers the pipeline.

Spec, code, art and copy in a single loop — fewer handoffs, and a PM who can answer technical questions with a running build.

Judgment gets scarcer, not cheaper.

Agents will build the wrong thing beautifully. Knowing which problem is worth solving is still the whole job.

Two stories

One product nobody had built before, and one that existed but had stopped working. Both came down to the same question: what is actually blocking the person on the other side of the screen?

AdColony · 2011–14 · VP Product Innovation

Nobody watches an ad they didn't ask for. So we let players ask for it.

The problem

Mobile video interrupted play. Publishers wouldn't run it because it cost them retention, and advertisers couldn't count on a completed view.

The call I made

Stop buying attention and trade for it. Value exchange for virtual currency, opt-in, built into the SDK — with Instant-Play video and HTML5 endcards so the wait that killed the experience disappeared.

What happened

$100M run-rate in three years and an acquisition by Opera Mediaworks. The demo apps I built for sales drove 300% revenue growth on their own.

What I owned: the SDK and API surface, analytics, dashboard, the creative studio ad builder, and the prototypes that got each of them approved.

$300M

Acquisition

Rewarded video went from an argument I was losing internally to the format the mobile industry runs on.

InMobi DSP · 2020–22 · Senior Product Manager

The demand was there. The console was the bottleneck.

The problem

Buyers spent their day fighting the platform instead of optimizing campaigns, and every new campaign waited on a design queue for creative.

The call I made

Treat creative, gaming and platform UX as one surface rather than three backlogs. Generate creative inside the product with Expresso, then clear console friction in shippable increments.

What happened

Revenue up 500% with performance metrics up across the board, and a top-5 global mobile DSP finish in 2022.

What I owned: the creative tooling roadmap, the gaming vertical and platform UX — a dozen-plus dashboard releases, plus two hackathon finals.

500%

Revenue growth

Most of it came from removing work buyers shouldn't have been doing, not from adding features.

Track record

2025 →

MAGLEV Games & Tech — Platform Innovation & AI

PULSE, a live ops and game backend platform. Product vision, architecture, pricing and positioning — anchored in a teardown of 14 competing platforms — through to global launch.

0→1, in build

2016 →

Superheart Studios — Founder

Cozy cross-platform games for mobile and PC — design, code, art direction and release. Built Heartbeat, the studio's own backend, and Tap Oceans: Pod Protector is in development.

Studio of one

2025–26

Stride Inc. via TiltedWindmill — Technical Product Manager

A new mobile app end to end: specs, sprints, third-party integrations, analytics, QA, App Store and Play releases.

3x retention lift

2016–24

AppOnboard / Buildbox — Product Lead, then Advisor

Playable ads and no-code game creation: iOS/Android SDKs, Studio, dashboard, Play Instant Apps, Snapchat playables beta. Back later to launch Buildbox 4 and research AI for the toolchain.

6 products launched

2008–11

Jirbo Games — Sr Producer, VP of Products

App Store launch day with 15 games, grown to 100+ titles. Built the beta ad SDK; platform PM for high scores, analytics and mediation.

50+ featured by Apple

University of Michigan, Ann Arbor · BS Political Science · Advisory through GLG and ExpertConnect · Full history on the résumé.

Built myself

The habit from step two, run at full size.

Nine things I specced, coded, art-directed and deployed on my own — agent-assisted, AI-generated art, no build step, live on a VPS. This is the part of the job most PMs hand off.

See more on GitHub

Engines, games, tools and sites — the rest of the shelf.

Bring me the messy platform problem.

Open to platform and web product roles, fractional PM, and build-with-you engagements. Westlake Village, CA · remote.

paulo.tb@gmail.com LinkedIn GitHub