Data platforms and lakehouses
A warehouse that doesn't need someone poking it every week to stay up.
Hi, I'm Bruno
Bruno Mendes
Data Engineering + AI Systems
Open to new projects · Book a call · Connect on LinkedIn
What actually ships
Data and automation that don't call you at 3am.
Pipelines, integrations, and AI agents running in production, with logs, retries, and monitoring wired in from day one. You find out something broke before your customer does — I've picked up enough 3am pages myself to build it any other way now.
The teams I build this for: finance, ops, marketing, growth.
What this avoids
- No silent pipeline breaks
- Every duplicate gets caught before it becomes rework
- You get alerted while there's still time to fix it
- Handoff without an emergency call
Live monitoring
- Integration Synced
- Next cycle in 1h
- Monitoring No issues detected
Daily pipeline
84%
128
runs / min
+30
clients served
GCP + AWS
delivery across cloud, warehouse, and activation stacks
Areas of work
The four things clients hire me for
Marketing and revenue integrations
Know which channel actually brought the customer, without stitching Meta Ads, GA4, and CRM together by hand to find out.
Finance and ops automation
No more reconciliation spreadsheet only one person knows how to open. ERP and finance talking directly.
AI workflows and signal pipelines
An AI agent that logs every decision it makes, so debugging it takes minutes, not days.
What I've actually built
Selected projects
Clean schemas, replay-safe pipelines, a handoff another engineer can pick up.
Finance automation
Financial Process Automation via Web Scraping
A Python scraping flow that logs in, reconciles what it finds, and turns finance's manual copy-paste routine into clean, BigQuery-ready records.
AI operations
AI Automation Engineer Workflows
Inbox ingestion, CRM task logic, and agent action records turn every automated decision into something with a paper trail, so you can pull up exactly why the agent did what it did.
Stack behind the projects above
- Python
- dbt
- BigQuery
- AWS
- GCP
- Lakehouse
- Dagster
- Observability
- AI Agents
- Workflow Automation
- Revenue Ops
- ERP Integration
- Finance Automation
- Marketing Attribution
- CRM Automation
- GitHub Actions
How an engagement runs
From first call to handoff
01
Scope
A short discovery pass, usually one or two calls, surfaces the handful of decisions that actually drive cost and architecture.
02
Build
The pipeline is running against real data within the first couple of weeks, and you can click into it and check for yourself.
03
Operate
Every pipeline ships with monitoring and safe retry paths built in from day one. I've cleaned up too many systems where that got added after the first outage.
04
Hand off
Runbooks and a repo structured so another engineer can extend it without a walkthrough call.