Skip to content

Analytics engineering · microsoft fabric

Christian Junior

Microsoft certified

Analytics Engineer | Fabric, lakehouse and Big Data.

When pipelines get heavy, numbers disagree with operations or leadership stops trusting the report, I rework ingestion, modeling and consumption. The goal is simple: data stops being overhead and becomes something you can decide on, with SQL, PySpark, notebooks, OneLake, Power BI and automation when the bottleneck isn’t “just ETL”.

View cases How I engage Contact
Professional portrait of Christian Junior

Where I usually show up

Solutions people call me for

The kind of thing that lands at 6pm or when the CFO asks “why did this number move?”. If the scenario sounds familiar, the next step is usually a technical conversation with a sponsor in the loop.

Capacity and load windows

Full loads filling a SKU (e.g. F16), queues backing up, shared workspaces suffering. I redesign incremental patterns, business keys, processing order and what it takes to reprocess without drama.

Numbers that don’t reconcile

P&L, accounting, project cost, HR: wrong grain or duplicated rules. I align sources, definitions and the semantic layer so BI stops being “a chart with opinions”.

Migration to Fabric

Moving off on-prem or Report Server without losing history or trust. Phased plan, workspaces, lineage and what becomes a new semantic model vs legacy.

Lakehouse turned swamp

Bronze without contracts, same entity in different shapes, consumption from the wrong layer. I tighten silver/gold, contracts and ownership: real Medallion, not just folders.

Semantics, Direct Lake and performance

Heavy models, long refresh or executive views inconsistent with operations. I decide what lives in the lake vs the model and where Direct Lake actually earns its keep.

Automation and capture

Processes living in email and spreadsheets, data born outside systems. Power Apps, Automate and lake landing zones when the bottleneck is human, not only transformation.

How I decide

Decisions and trade-offs you see in the real world

Short notes from the kind of conversations I have with architects, business and operations, no “innovation” slide deck.

  • Incremental instead of full

    When the window or SKU can’t afford to reload everything every time, the design shifts: watermark, business key, idempotency where needed and reprocess testing.

    Trade-off: more operational and test complexity in exchange for stable capacity.

  • Direct Lake in one place, import in another

    Direct Lake when model latency and scale matter; import when DAX/transform flexibility still wins over refresh pain.

    Trade-off: tighter coupling to the lake vs freedom inside the model.

  • Minimum viable governance

    Without metric owners, folder standards and promotion rules between layers, “gold” becomes a shared drive and BI becomes an endless debate.

    Trade-off: a bit of process in exchange for one official number.

  • BI tied to a decision ritual

    A report without committee, owner or cadence becomes wallpaper. I structure KPIs and views to support a meeting that already exists, or kill a meeting that only existed for lack of data.

    Trade-off: less “creative layout” freedom, more adoption and use.

  • Contract in silver, not hope in bronze

    Bronze lands; silver standardizes and tests; gold answers the business question. Skip a layer and you pay later in reconciliation and BI rework.

    Trade-off: longer path to the first pretty SELECT in exchange for fewer surprises later.

  • Engineering speaking the same language as the decision

    I translate technical constraints into impact (time, risk, cost) and pull the business into grain decisions early, so nobody “signs off” a BI artifact they don’t actually own.

    Trade-off: tougher meetings up front in exchange for less rework at the end.

About

I don’t sell a pretty dashboard. I sell clarity under pressure.

I work hands-on in the Microsoft analytics stack: lakehouse, layers, incremental loads, semantics, governance and consumption. I’ve been in the room for Fabric capacity pain, migrations off on-prem / Report Server, and models that only behaved once grain and layers were right.

What shows up in a meeting may be a report; what sustains a decision is data contracts, pipeline cost, metric definition and who runs it after the project “ends”. That’s why I document and standardize: analytics without operations becomes shelfware.

In hybrid teams I often sit between business, platform and IT: translate real constraints (window, SKU, policy) into a design that doesn’t depend on a hero engineer.

  • Lakehouse on Fabric/OneLake: Medallion with explicit rules per layer, no “read straight from bronze because it was faster”.
  • Semantics & BI: SQL, DAX and Direct Lake when the issue is latency, trust or financial close, not just visuals.
  • Power Platform when data is born wrong or the process is manual: Apps, Automate and lake destinations when it fits.
Full journey

Cases

Real problem, decision, delivery and what changed for the business

Each write-up is a closed narrative: context, what was breaking (capacity, trust, time), architecture choices, stack and impact. BI shows up when it supports a committee, P&L, ops or HR, not as a lonely screenshot.

Cases are being published. The full list is coming soon.

Solutions

How I join a project

Depends on the stage: sometimes pipeline triage, sometimes lakehouse design, sometimes executive BI with an official metric. The thread is closing the loop between engineering and decision, with something your team can run.

Data structuring for decision-making

I turn business questions into reliable metrics, the right grain, and an analytical narrative that supports executive committees.

Deliverables

  • KPI map and initial dictionary
  • Data model / measure prototype
Solution details →

Pipeline engineering and data ingestion

Incremental pipelines, source contracts, and minimum observability: Fabric, notebooks, or equivalent stack.

Deliverables

  • Pipeline and dependency design
  • Implementation and operational documentation
Solution details →

Lakehouse implementation on Microsoft Fabric

OneLake, domain-based lakehouse structure, PySpark notebooks, and BI publishing with lightweight, repeatable governance.

Deliverables

  • Lakehouse blueprint and conventions
  • First end-to-end domain (MVP)
Solution details →

Executive and operational dashboards

Dashboards tied to real decisions: visual hierarchy, agreed KPIs, and performance that supports daily use.

Deliverables

  • Validated prototype + semantic model
  • Publishing with performance standards
Solution details →
Scopes & details

Path

Path

From IT operations to analytics engineering: infrastructure, BI, projects and now Fabric/lakehouse with the business problem in the center.

  1. Analytics Engineer | Data & AI

    Inmetrics March 2026 · present

    São Paulo, Brasil

    End-to-end analytics architecture on Microsoft Fabric, from ingestion and lakehouse design to executive consumption with BI and applied AI.

    • Fabric pipeline structuring focused on governance, performance, and reliability
    • Notebook and PySpark usage for higher data volumes and complex rules
    • Connecting engineering, analytics, and executive decision-making in business-oriented solutions

  2. SR IT Project Analyst | Business Intelligence

    Cast group May 2025 · March 2026

    São Paulo, Brasil

    Leading BI and analytics projects focused on metric standardization, architectural evolution, and predictable delivery.

    • Planning and executing analytical streams with multiple stakeholders
    • Structuring data processes to improve quality and governance of indicators
    • Improving executive consumption layer with focus on decisions, not screen volume

  3. IT Project Analyst | Business Intelligence

    Cast group May 2024 · May 2025

    São Paulo, Brasil

    BI initiatives with project perspective, integrating business needs, analytical modeling, and delivery quality.

    • Evolution of reporting flows into more reliable analytical models
    • Technical alignment between business, data, and technology
    • Support in backlog organization, prioritization, and metric governance

  4. Business Intelligence Analyst

    Cast group February 2023 · April 2024

    São Paulo, Brasil

    Outsourcing engagement for Sao Paulo State Finance and Planning Department, focused on analytical reliability and public management support.

    • Data and indicator consolidation for executive monitoring
    • Support in structuring consistent metrics across teams
    • Deliveries focused on clarity, context, and decision support

  5. Business Intelligence Analyst

    Grupo Vamos January 2022 · February 2023

    Mogi das Cruzes, São Paulo, Brasil

    Corporate BI role connecting operational performance to business indicators for decision-making.

    • Structuring reports and indicators for operations and management
    • Improving data quality to reduce analytical rework
    • Enhancing executive readability with KPI-to-action focus

  6. Performance Analyst

    Grupo Vamos December 2021 · March 2022

    Mogi das Cruzes, São Paulo, Brasil

    Transition from operational support to performance analysis, focused on KPI monitoring and process efficiency.

    • Performance analyses supporting short-term decisions
    • Data organization and systematic KPI tracking

  7. Entry-Level IT Infrastructure Analyst

    Neobpo November 2021 · December 2021

    Mogi das Cruzes, São Paulo, Brasil

    Work in infrastructure and technical support focused on operational continuity, foundation for later growth in data and analytics.

    • Technical support for corporate environment
    • Support for infrastructure routines and operational stability

  8. Entry-Level IT Assistant

    Neobpo January 2021 · November 2021

    Mogi das Cruzes, São Paulo, Brasil

    Beginning of my technology journey, with operational support and technical routines that built process discipline and quality.

    • User support and incident resolution for low and medium complexity issues
    • Process follow-up and basic technical documentation

Credentials

Certifications and certificates

Certification doesn’t replace delivery, but it shows I play by platform rules. Highlight on what backs serious BI; the rest is grouped as Microsoft, data and methods.

Badge da certificação Data Science Orientation

Data & Analytics

Data Science Orientation

IBM

2024

Badge da certificação ITIL® Foundation Certificate in IT Service Management

Governance & Methods

ITIL® Foundation Certificate in IT Service Management

Código da credencial GR671531620CD

2023

All credentials & official links

Stack

Where I work

Stack organized by workstream.

Data engineering

Microsoft Fabric OneLake & lakehouse architecture Data Factory & ETL/ELT pipelines PySpark & notebooks SQL & incremental ingestion Layered modeling (Medallion)

Analytics & BI

Power BI Semantic modeling DAX KPI framework & storytelling Executive dashboards for decisions

Automation & apps

Power Automate Power Apps AI Builder Integrations & workflows Process-oriented automation

AI & applied intelligence

Fabric ML Analytical models AI applied to the business Enriched processes and decisions

Governance & reliability

RLS & access security Workspace organization Visual & semantic standards Analytical reliability Compliance & operational documentation

Fabric, sensitive data or a migration that can’t fail?

Send one sentence on the problem (e.g. full loads killing capacity, CFO mistrust in the P&L, leaving Report Server). I’ll reply with what I’d do first and what I wouldn’t promise.

Open contact LinkedIn