BACKEND SYSTEMS / DATA ENGINEERING

Dênio Barbosa
Júnior.

Software Engineer Backend & Data Systems

I build and improve systems where reliable data matters. My work spans banking applications, data validation, and file-based integrations using SQL, Db2, COBOL/CICS, and C#/.NET.

01 / ABOUT

Engineering with
care for the details.

I’m a software engineer focused on the path from a business requirement to dependable data and software behavior.

In banking systems, I’ve worked on batch processing, historical data management, and investigations that require technical precision and collaboration across teams. I also explore modern data workflows in personal projects.

My next step is to contribute to backend systems, data engineering, or analytics engineering work where correctness and clear thinking have lasting value.

Current work
Software Engineer (Contract), Banco Mercantil do Brasil
Focus
Backend systems · Data validation · Integrations
Languages
Portuguese (native) · English (professional)

02 / SELECTED WORK

From sources to
usable information.

A personal ETL project that connects ingestion, transformation, orchestration, and exploration in one workflow.

PERSONAL PROJECT / DATA PIPELINE

Brazilian Financial
Markets ETL Pipeline

An end-to-end project exploring how Brazilian financial market data moves from ingestion into a modeled PostgreSQL warehouse and a Streamlit interface. Developed with AI assistance; the case study separates the repository’s implemented pieces from the intended architecture.

PythonPostgreSQLdbtAirflowDockerStreamlit

03 / EXPERIENCE

Systems that hold
up in practice.

One continuous banking engagement, with a change in contracting arrangement in October 2024.

APR 2024 — PRESENT

Software Engineer (Contract)

Banco Mercantil do Brasil

Contracted through IEBT Innovation until October 2024; direct contractor thereafter.

  • Historical data at scale. Designed and deployed a Db2 archival approach moving 190 million historical log records into a history table, supported by monthly batch processing.
  • Validated imports. Designed a file-driven association import and implemented a COBOL batch program that validates approximately 200 partner-product associations per run.
  • Data-quality investigation. SQL analysis surfaced 847 initially incomplete upstream records; a cross-team investigation identified more than 3,000 affected records across the broader issue.

04 / INTERACTIVE LAB

See data
decisions unfold.

Explore two browser-only demonstrations: validate a sample batch, or filter an invented market-like series and see every chart recalculate.

Both labs use synthetic data created for this website. They do not use employer, customer, live market, or project data.

05 / ARCHITECTURE

How the work
fits together.

A closer look at design decisions, with conceptual illustrations where implementation details are confidential.

01Market data ETL pipelinePersonal project · source → ingestion → modeling → exploration

The project organizes ingestion with Python, relational storage in PostgreSQL, transformations with dbt, scheduling with Airflow, and a Streamlit exploration layer. The case study describes the repository and its current limits in more detail.

Project architecture. The diagram is an explanatory map, not evidence of a live deployed service.

02Validated file-driven importProfessional experience · conceptual pattern

A batch import can validate a file’s required fields and associations before updating downstream records, returning actionable exceptions rather than leaving manual one-at-a-time entry as the only path. The public description is intentionally generalized from my work on a COBOL batch program.

Sanitized concept. No employer schema, code, or partner data is shown.

03Historical Db2 archivalProfessional experience · conceptual pattern

Historical records are separated into a history table through a monthly process, keeping the distinction between current operational data and retained history explicit. My work involved designing and deploying an approach that moved 190 million historical log records.

Sanitized concept. Scale is a documented work result; internal design details are omitted.

04Website validation workbenchOriginal demonstration · synthetic records

The interactive lab applies explicit required-field, date, and duplicate checks to deterministic synthetic records. Filters change the view, while validation order and record identities remain stable.

Educational website demonstration. It is separate from my employer systems and market-data project.

06 / CAPABILITIES

Tools in their
working context.

Grouped by where I have used or explored them, rather than a made-up proficiency score.

01 / PROFESSIONAL SYSTEMS

Banking and backend

Batch processing, data investigation, and application work in a production banking environment.

  • SQL
  • Db2
  • COBOL / CICS
  • C# / .NET
  • File integrations
02 / DELIVERY

From requirement to run

Designing validation paths, tracing data issues, and coordinating findings with other teams.

  • Data validation
  • Batch design
  • Problem analysis
  • Cross-team investigation
03 / PERSONAL PROJECT

Modern data workflow

Tools used in the Brazilian Financial Markets ETL Pipeline, a personal project developed with AI assistance.

  • Python
  • PostgreSQL
  • dbt
  • Airflow
  • Docker
  • Streamlit

07 / EDUCATION & RÉSUMÉS

A foundation for
what comes next.

2022 — 2023 / TORONTO

Software Engineering Technician

Ontario College Diploma · Centennial College

Choose a résumé focus

Each version presents the same professional history for a different recruiting context.

08 / CONTACT

Let’s build something
dependable.

Interested in backend systems, data engineering, or a technically demanding problem? I’d be glad to talk.