Data Engineer and Data Analyst with 15 years of experience across the whole data lifecycle: ingestion, transformation, modeling, governance and visualization.
- 🏗️ I build data platforms from scratch in companies with little or no analytics culture, implementing medallion architecture (bronze, silver, gold), dimensional models and governed data products for BI.
- ☁️ Hands-on with modern cloud stacks: AWS (S3, Glue, Athena), Snowflake, Azure (Data Factory, Synapse) and Databricks, including cross-cloud migrations (AWS → Snowflake on Azure) and cost optimization.
- 📊 Expert in SQL, Python and Power BI, turning data into reliable insights for executive decision-making — including enterprise Power BI practices: PBIP/TMDL in Git, row-level security, incremental refresh and calculation groups.
- 🔄 Data migrations between systems: source-to-target mapping, data cleaning rules, full and incremental loads and reconciliation that proves nothing was lost.
- 📈 Statistics applied to business questions: hypothesis testing, confidence intervals, experiment design (A/B, champion vs challenger) and credit risk scoring.
- 🤝 Delivered data projects for companies such as Itaú, JBS (Seara, Friboi, Swift), Suhai Seguradora, Samsung, Panasonic and Nestlé.
- 🌏 Advanced English — 2 years living and working in Australia.
Languages
Cloud & Data Platforms
Data Engineering
Analytics & Statistics
BI & Visualization
Databases (SQL & NoSQL)
| Project | Description | Stack |
|---|---|---|
| Fintech Credit Risk Analytics | Credit risk analytics on 307k real loans and 17M credit history records: Databricks medallion pipeline, hypothesis tests on business rules, credit scorecard (AUC 0.755) and approval cutoff simulation — the score cuts defaults by 42% vs the current rules at the same approval rate | Databricks · PySpark · Spark SQL · Delta Lake · MLlib · SciPy |
| Legacy Banking Data Migration | Migration of a legacy core banking database (SQL Server) to PostgreSQL and MongoDB: source-to-target mapping and cleaning rules, full and incremental loads, reconciliation gate (every transaction compared field by field, money totals and app balances), rejected records with reasons, alerts, a FastAPI service and a Power BI cut-over dashboard (backlog burn-down, reconciliation history) — CI on real SQL Server, PostgreSQL and MongoDB containers | Python · SQL Server · PostgreSQL · MongoDB · SQLAlchemy · FastAPI · Power BI · Docker |
| Auto Insurance Data Warehouse | Restructuring of a slow legacy SQL Server for an auto insurer: ODS and star-schema data warehouse (SCD2 customer address, monthly earned premium snapshot, claim lifecycle), incremental T-SQL loads in a single transaction, a release gate that reconciles premium and claims with the legacy system to the cent, SQL Server Agent scheduling and a Power BI dashboard (loss ratio, brokers, claims and theft) — the warehouse returns exactly the legacy loss ratio report with 608× fewer reads; CI on SQL Server 2022 | SQL Server · T-SQL · Columnstore · SQL Server Agent · Power BI · Python |
| Public Spending Data Platform | AWS lakehouse migration to Snowflake: ingestion, Glue, Athena, dbt star schema with data quality tests, shadow validation (99.5% gate), monthly Airflow orchestration, cost analysis, Terraform and CI | AWS S3 · Glue · Athena · Snowflake · dbt · Airflow · Terraform |
| E-commerce Product Analytics on BigQuery | Product and marketing analytics on the real GA4 export of the Google Merchandise Store: dbt models over nested events, purchase funnel, channel conversion with confidence intervals and significance tests, weekly cohort retention, RFM segmentation and a public Looker Studio dashboard | BigQuery · dbt · SQL · Python · Looker Studio · GitHub Actions |
| E-commerce Logistics Data Platform on Azure | Delivery and freight analytics for ~100k orders: Data Factory orchestration, ADLS Gen2 medallion lake, Synapse Spark and serverless SQL, geospatial route analysis and a data quality gate | Azure Data Factory · Synapse · ADLS Gen2 · PySpark |
| Contact Center SLA — Power BI | Enterprise Power BI project for outsourced customer service (6 BPO partners, voice and WhatsApp, 1.5M contacts): PBIP/TMDL in Git, contract-based SLA measures, calculation group for time intelligence, dynamic row-level security, incremental refresh, report as code and a CI quality gate with the Tabular Object Model and 19 best-practice rules | Power BI · DAX · Power Query · TMDL · Python · GitHub Actions |
| Data Lakehouse on Databricks | End-to-end lakehouse with medallion architecture, config-driven ingestion, data quality rules and star schema | Databricks · PySpark · Delta Lake · Spark SQL |
| Public Budget Execution Dashboard | Analysis of Brazilian public spending with official government data: execution, regional distribution and KPIs | Power BI · DAX · Power Query |
- Data architecture: lakehouse and data warehouse design, ODS, medallion layers, star schema
- Pipelines & ETL: batch ingestion, orchestration, Glue/ADF/Databricks jobs, SQL-based transformations
- Data migrations: source-to-target mapping, cleaning rules, incremental sync and reconciliation gates before cut-over
- Data quality & governance: validation with business areas in staging, data and process testing, monitoring of daily routines
- Cost & performance: query optimization, workload migration to pay-per-use compute
- Advanced analytics: hypothesis testing, experiment design, credit scoring and policy simulation in money terms
- BI & storytelling: executive dashboards and governed self-service models in Power BI (RLS, incremental refresh, calculation groups, CI checks)