Juliana

Analytics Engineer · Remote · Brazil (GMT-3)

I turn scattered data into pipelines the business trusts — and decisions that hold up.

I'm Juliana, an Analytics Engineer. I model data in dbt, build reliability in with automated tests, and turn all of it into decisions business teams can act on without asking me twice.

View case studiesGet in touch
SQLdbtPythonBigQuerySnowflakeAirflowLooker

How data goes from scattered to trustworthy — click a stage.

Product events, business spreadsheets, and third-party APIs — each with its own format and its own version of the truth.

APIsCSV/SheetsWebhooks

3

case studies documented end to end

4

stack layers, from staging to BI

100%

of fact models covered by automated tests

GMT-3

remote, Brazil timezone

About

I work at the intersection of data engineering and the questions the business actually needs answered. My work starts where the data is still raw and ends when someone can make a decision without having to ask me if the number is right.

That means careful modeling, automated tests as the default rather than an extra, and documentation that outlives my own memory. I'd rather ship a simpler, more reliable pipeline than a cleverer, more fragile one.

Outside the warehouse, I like sketching how information should flow before writing the first line of SQL — most of the data bugs I've fixed started as a modeling problem, not a code problem.

Stack

4 layers

Modeling & Transformation

where raw data becomes something you can trust

dbtAdvanced SQLDimensional modelingPython (pandas)

Warehouse & Orchestration

where pipelines run and get observed

BigQuerySnowflakeAirflowdbt Cloud / CI

Quality & Reliability

what keeps a wrong number from reaching the dashboard

dbt testsGreat ExpectationsData contractsLiving documentation

BI & Communication

where analysis becomes a decision for people who don't write SQL

LookerMetabasePower BIData storytelling

Case studies

Impact metrics below are illustrative — the real projects exist, but the final numbers are still being added here.

01

One single source of truth for revenue

Three teams were reporting different revenue numbers for the same question. The work was to model until only one right answer was left.

Problem

Finance, growth, and product each calculated “monthly revenue” in subtly different ways — parallel spreadsheets, manual joins, different rounding. Every meeting started with ten minutes spent agreeing on whose number was correct.

Approach

Dimensional modeling from scratch in dbt: one staging layer per source, a single fact model for revenue with documented business rules, and automated tests covering cases that had silently broken before.

Reconciliation between teams
from days to minutes
Divergent revenue sources
3 → 1
Test coverage on the fact model
100%
dbtBigQuerySQLdbt tests
02

Observability for pipelines no one could see break

Silent failures reached the executive dashboard before they reached me. The fix was to prevent, not to hunt.

Problem

Critical pipelines were failing partially — they didn't crash, they just delivered incomplete data. The first signal was usually someone from leadership asking why a number had dropped by half.

Approach

Data contracts between producing and consuming teams, anomaly and volume tests at every critical stage, and alerts wired directly to the responsible team's channel — before the data ever reached the BI layer.

Failures caught before BI
nearly all
Time to identify root cause
consistently reduced
Critical pipelines under contract
full coverage
Airflowdbt testsGreat ExpectationsSlack alerts
03

A warehouse the business team could use on its own

Every new question turned into a ticket for me. The fix was modeling for autonomy, not just analysis.

Problem

Business teams depended on an analyst for any question outside the ready-made dashboards — even simple questions turned into a days-long queue.

Approach

A documented semantic layer using business (not database) naming, core metrics defined exactly once, and short onboarding sessions for each team to explore the data safely.

Ad-hoc analysis tickets
sharp drop
Teams exploring data unassisted
most
Core metrics with a single definition
yes
Lookerdbt docsSQLBigQuery

Experience

2023 — present

Analytics Engineer · [Company]

Responsible for the warehouse's transformation layer: dbt modeling, quality tests as the team standard, and direct support for the product and growth squads.

2021 — 2023

Senior Data Analyst · [Company]

Bridge between the data team and business areas. Gradual migration from spreadsheets to versioned, documented warehouse models.

2019 — 2021

BI Analyst · [Company]

Built the company's first standardized dashboards and defined the core metrics the team still uses today.

Quick questions

For anyone with thirty seconds before the next interview.

Analytics Engineer or senior Data Analyst with a modeling focus — roles where someone needs to turn raw data into a reliable, documented warehouse, not just write one-off queries.

Channel open

Want to talk about the next pipeline?

Open to opportunities as an Analytics Engineer. If you want to talk about a role, a project, or just swap notes on data, the channel is open.

LinkedIn GitHubResume