Senior PM who owns products end-to-end with data rigor — from 0-to-1 build to experimentation infra to cross-functional discovery — currently applied to pricing & monetization at Brazil's largest bus marketplace.

The problem

ClickBus' low-fare product — advance purchase of carrier seat inventory resold with dynamic fares — was relevant but couldn't scale: billing was 100% manual, margin guardrails lived in spreadsheets with no real-time integration, discounts were fixed regardless of context, campaign management was fragmented, there was zero experimentation infrastructure, and search responded in ~400ms. Roughly 9% of purchased seats expired unsold.

What I did

• Shipped a unified campaign-management platform: parameterized search, 360° campaign view, one-click activation via API.
• Built a proposal simulator & validator that automated commercial proposal generation and post-negotiation financial validation — replacing an external tool dependency.
• Launched dynamic per-bus pricing ("smart discount") on a 3-layer architecture deliberately designed to evolve from decision matrix to ML without a rewrite.
• Automated billing end to end (balance write-off, mark-up, guardrails, over-sale) and moved margin guardrails from spreadsheets to real-time calculation in the database.
• Stood up permanent experimentation infrastructure (Unleash + Amplitude) and ran the company's first incrementality A/B test.

The problem

The low-fare flow historically cancelled 4–5 percentage points above the regular flow — a persistent revenue leak nobody had root-caused.

What I did

• Ran my own data analysis (Python, data lake) and identified portfolio mix as the driver: low-fare tickets lacked the exchange/cancellation self-service the regular flow had.
• Redesigned the exchange and cancellation flows to support low-fare tickets, shipped in two waves.
• Measured impact with a counterfactual analysis against the historical spread — presented at c-level review.

The problem

Leadership debated whether low fares actually created demand or just cannibalized full-price sales — with no experimental evidence either way, and no infrastructure to get it.

What I did

• Designed the experiment end to end — 17 routes of a major partner carrier, 8,400+ users, deterministic bucketing, pre-defined significance criteria.
• Built the permanent experimentation stack (Unleash + Amplitude) so every future test reuses the same infrastructure.
• Documented the results as a negotiation asset now used by the commercial team with carriers.

The problem

Pricing decisions across ClickBus' sales channels were made with a fragmented view — marketing, strategy, commercial, product and engineering each held a partial picture. The relationship with partner carriers lived in silos too: commercial negotiated inventory, finance fought processes case by case, and nobody owned a unified read on the B2B experience.

What I did

• Ran structured interviews with stakeholders across 5 internal areas (marketing, strategy, commercial, product, engineering) to build a unified channel pricing strategy — in progress.
• Ran recurring discovery with partner carriers across 3 distinct workstreams: inventory negotiation, financial process mapping, and B2B experience design.
• Inventory-negotiation discovery fed directly into the company's largest seat-inventory deal in recent years.
• Financial-process discovery surfaced the validation gaps later formalized during the product's external audit.
• B2B-experience discovery led tribe leadership to invite me to co-design the company's B2B product, its main goal for the year.

The problem

One PM, four products: keeping the delivery bar high across discovery, specs, GTM and impact analysis doesn't scale by working more hours. I rebuilt the way the work itself gets done.

What I did

• Built a pipeline of 14 custom Claude Code skills covering the full cycle: raw data → analysis → discovery brief → PRD → prototype → spec + ADRs (with automated validation) → issues → implementation → GTM → impact analysis.
• Automated board governance with GitHub Actions (board health, epic→feature hierarchy, PR insights) and migrated squad tracking from Jira to GitHub Projects.
• Reverse-engineered a legacy pricing microservice into full documentation — C4 diagrams, ERD, state machines, OpenAPI, 7 ADRs — using an AI-assisted framework.
• Mentor PMs and engineers across the company; share cases at tribe meetings with the stated goal of raising the whole product team's bar.