Jihwan Kim, Power BI & Microsoft Fabric Architect

About me

I architect the platform. I also build the tooling for it.

Enterprise teams everywhere are working through the same transitions right now: Fabric migration, PBIR adoption, TMDL-based CI/CD, and semantic models that Copilot and AI agents can answer from, all without breaking production reports. This is the work I do every day.

I started in supply chain. At Samsung SDS Cello Logistics in the Netherlands I managed key accounts across global freight and warehouse operations and discovered what a single Power BI dashboard could do to a report-driven operations culture. That pivot stuck.

Since then I have built BI for the operators I used to be: GXO Logistics across 10+ EU sites, ITVT Group on Dynamics 365 supply chain, and now IKEA Customer Fulfilment, Inventory & Logistics. At IKEA I architected the migration from Import mode to Direct Lake on Microsoft Fabric for 30,000+ users and billions of rows. It removed over 90% of the refresh load, made the models 30% smaller, and sped up queries by 25%.

I also ship the DevOps tooling that makes the Fabric/PBIR/TMDL transition repeatable. Five open-source tools (one Python, four JavaScript), 95+ GitHub stars, used by Power BI teams in other companies. As a Microsoft MVP I test PBIR and DAX UDF features with the product group before they ship.

And I lead how my team uses AI in everyday development. The tooling moves fast, so I keep testing what is new, from Claude to Rayfin CLI and Fabric Apps, and fold the parts that work into how we build. Right now I am learning and actively using the Power BI Desktop Bridge and the Skills for Fabric agent skills: the agent edits the PBIR and TMDL files on disk, then reloads and verifies against the live Power BI Desktop session. That practice has roughly tripled our development speed, and I coach the analysts to do the same.

Microsoft MVP since 2022, recognized for community contributions in Power BI and Fabric: direct access to Microsoft's roadmap, early previews of PBIR and DAX UDFs, and a direct feedback channel to the engineering team.

Community Super User since 2021, 2,700+ solutions authored in the Power BI & Fabric forums. Active community mentor.

Users served

30,000+

supply chain

Refresh time reduced

90%+

Direct Lake

Model size reduction

30%

optimized

Performance gain

25%

CU savings

Open-source tools

95+ stars

Years Power BI

enterprise

Open source

Tools enterprises need for PBIR/TMDL governance.

The DevOps tooling I built from real-world pain at IKEA scale, free and open-source. 95+ GitHub stars · 5 tools · solving PBIP/TMDL/PBIR governance worldwide.

pbip-documenter star41 · July 2026

Generates documentation from Power BI semantic models. Reads PBIP/TMDL files and pulls every measure, table, relationship, and metadata entry into readable output. Used by enterprise teams managing 100+ semantic models.

JavaScript PBIP Documentation

View on GitHub open_in_new

Problem

BI teams spent hours documenting semantic models by hand. The documentation drifted from reality, and new team members had nothing current to read.

Approach

Built a JavaScript parser that reads PBIP/TMDL project files and extracts every measure, table, relationship, and metadata entry into readable documentation.

Impact

Documentation time reduced from hours to seconds. Adopted by Power BI teams across multiple organizations. 41 GitHub stars and growing.

pbip-lineage-explorer star35· July 2026

Traces DAX measure dependencies from visuals to source columns in PBIP projects. Includes commit-by-commit change intelligence across 30+ change types. Essential for teams adopting Git-based workflows.

JavaScript DAX Lineage Git Change Intelligence PBIP

View on GitHub open_in_new

Problem

DAX refactoring in large PBIP projects was blind, no way to know which visuals would break when renaming or removing a column or measure.

Approach

Built a dependency tracer that maps DAX lineage from visuals to source columns, plus commit-by-commit git intelligence across 30+ change types.

Impact

Teams refactor with full visibility into downstream effects. Change intelligence eliminates guesswork in code reviews.

isHiddenInViewMode star10· July 2026

Bulk-manage filter visibility and layer order in PBIR reports. Eliminates tedious manual clicking through Edit Interactions. The first tool to enable bulk management of PBIR visual properties.

JavaScript PBIR Productivity

View on GitHub open_in_new

Problem

Managing filter visibility in complex PBIR reports meant clicking through each visual one by one, tedious and error-prone for reports with dozens of visuals.

Approach

A bulk manager that reads PBIR JSON directly, enabling batch operations on filter visibility, layer order, and visual interactions.

Impact

What took 30+ minutes of repetitive clicking now takes seconds.

pbip-impact-analyzer star5· July 2026

Analyze impact of changes and safely refactor PBIP semantic models. Know which visuals, measures, and reports are affected before you make the change. The safety net for enterprise refactoring.

JavaScript Impact Analysis Refactoring

View on GitHub open_in_new

Problem

Renaming or removing measures was risky, no way to predict which reports, visuals, or downstream measures would break.

Approach

Scans the entire PBIP project to map dependencies before any change, showing exactly which visuals, measures, and reports are affected.

Impact

Teams refactor with confidence. Safe refactoring becomes the default workflow.

pbip_model_lenz star3· July 2026

PBIP analyzer that bridges BI developers and data engineers, one graph, two readings. Click any measure to trace every table it depends on through direct references and active relationships. First Python tool in the suite, shipped on PyPI as pip install model-lenz.

Python PBIP Lineage PyPI

View on GitHub open_in_new PyPI open_in_new

Problem

BI developers and data engineers read the same semantic model with different vocabularies. BI sees measures and PBIP table names; data engineers see source-system tables they own. Impact analysis at the handoff slows down because nobody owns the translation.

Approach

Static analysis of PBIP metadata builds one dependency graph that can be rendered in either vocabulary. Click any measure to see every table it really depends on, including dependencies reached through active relationships, not just direct references. Distributed as a Python package on PyPI so it drops into CI without per-machine setup.

Impact

Cross-functional teams stop translating between vocabularies in review meetings. Data engineers see the source-system names they recognize; BI developers see the measure semantics they own. Same graph, both audiences.

Career

Enterprise-scale impact.

I spent over ten years in supply chain and logistics operations before moving into BI. That background is why I design around the numbers operations teams actually act on, not just the model behind them.

IKEA

CURRENT

Power BI Architect

Feb 2023, Present · Amsterdam, Netherlands

Scope: Customer Fulfilment, Inventory & Logistics Insights, global supply chain analytics

Architect semantic models and analytics pipelines on Microsoft Fabric and Power BI serving 30,000+ users across the global supply chain organization

Key achievements

Semantic model development process optimization

Implemented systematic parameterization, source-control, row-count, and data-integrity parameters, making development controllable and reproducible across environments. Reproducible deployments across 3 environments, zero deployment failures, release time cut 70%.

Git integration & source-control architecture

Designed a new architecture achieving full source control and version control for PBIP/TMDL artifacts, with CI/CD pipelines via Azure DevOps and GitHub. Zero-downtime deployments, full audit trail for compliance, 50% faster onboarding.

Direct Lake & composite model migration

Migrated legacy Import-mode models to composite models with Direct Lake. Refresh load dropped by more than 90%, CU usage came down, and users got more data with 25% faster queries. Near real-time analytics for 30,000+ users; data latency from hours to minutes.

Reworked the semantic models for a 30% size reduction and a 25% performance gain
Reduced Fabric Premium CU usage by 25% through hands-on tuning and AI-assisted performance analysis
Integrated AI copilots into DAX authoring, performance diagnostics, and documentation, now learning and actively using the Power BI Desktop Bridge and Skills for Fabric agent skills, boosting dev speed 3×
Mentored analysts on semantic modeling, DAX optimization, and AI-powered report generation

ITVT Group

Senior BI Analyst

Jun 2022, Jan 2023 · Utrecht, Netherlands

Delivered Power BI solutions across Supply Chain, Logistics, CRM, HR, and Project Management, integrated with Dynamics 365
Built a logistics performance dashboard consolidating shipping, warehousing & fulfillment KPIs, weekly reporting from 6 hours to 10 minutes
Optimized query folding and DAX logic, improving load times by 40–60%
Built reusable Power BI templates and refactored legacy BI components for faster client rollouts

GXO Logistics

Business Intelligence Analyst

Jun 2021, May 2022 · Schiphol-Rijk, Netherlands

Designed and standardized Power BI dataflows, datasets, and reports across 10+ EU sites, Dock-to-Stock, On-Time Shipping, Inventory KPIs
Delivered Weekly Stability, Monthly Business Review, and Productivity & Staff Planning dashboards
Delivered near real-time insights to site controllers and site management for forecasting, benchmarking, and weekly planning

Supply Chain Flow

Supply Chain Data Analyst (Freelance)

Feb 2020, May 2021 · Amsterdam, Netherlands

Built Power BI dashboards for logistics and warehouse teams covering freight pricing, bottleneck analysis, and cash flow visibility
Launched FlowAnalysis.co, deep dives on data modeling, DAX patterns, and supply chain analytics
Earned APICS CSCP, MCSA: BI Reporting, MCT, and MOS Excel Expert

Samsung SDS Cello Logistics

Business Development Manager

Feb 2019, Feb 2020 · Delft, Netherlands

Career pivot from logistics operations into data analytics at Samsung SDS Cello, discovered Power BI and what data could do for supply chain decisions
Led key account freight and warehouse operations and introduced data-driven reporting into a report-driven culture

Capabilities

What I bring to the platform.

Four areas I work in every day. The numbers are self-rated depth, grounded in the outcomes above.

analyticsSemantic model architecture

Import, DirectQuery, Direct Lake, and composite models at billion-row scale. Physical & logical data modeling, role-playing dimensions, incremental refresh, aggregations, object-level security.

Proof: IKEA 30K users · 90%+ refresh elimination

Direct LakeComposite modelsTMDLOLS/RLS

Depth95 / 100

account_treeDevOps & governance (PBIP/TMDL)

Git-based source control for Power BI artifacts, Azure DevOps / GitHub Actions pipelines, PBIP/TMDL workflows, automated documentation, impact analysis, and environment promotion.

Proof: IKEA CI/CD · 70% faster releases · zero-downtime deploys

PBIPAzure DevOpsGitHub ActionsPBIR

Depth92 / 100

buildDAX & performance engineering

Time intelligence, advanced calc groups, query plan analysis, VertiPaq tuning, DAX Studio, Tabular Editor scripting, UDFs, and calculation group design for enterprise semantic layers.

Proof: 2,700+ community solutions · 25% query performance gain

DAX StudioTabular EditorVertiPaqCalc groups

Depth94 / 100

boltAgentic BI engineering

AI agents in every phase of development: Claude for DAX authoring, PBIR/TMDL editing, documentation, and performance diagnostics, and now the Power BI Desktop Bridge and Skills for Fabric agent skills, which I am learning and actively using so agents edit the files on disk and verify against the live Desktop session. Semantic models designed so Copilot and AI agents answer correctly. I test each new tool as it lands and coach the team on what works.

Proof: 3× development speed at IKEA · team-wide adoption

ClaudeDesktop BridgeSkills for FabricCopilotAI-ready models

Depth90 / 100

Microsoft BI Platform

Power BI (Desktop & Service)8+ yrs

Microsoft Fabric2+ yrs

DAX8+ yrs

Power Query / M8+ yrs

TMDL / Tabular Editor4+ yrs

Data & DevOps

Azure DevOps / GitHub Actions3+ yrs

PBIP / PBIR2+ yrs

SQL (T-SQL)7+ yrs

Python (DAX tooling, analysis)3+ yrs

Direct Lake / OneLake2+ yrs

Dynamics 365 integration1+ yrs

Recognition

Achievements & certifications.

Certifications

Active

PL-300: Power BI Data Analyst DP-600: Fabric Analytics Engineer AZ-900: Azure Fundamentals Power Platform Fundamentals MOS: Excel Expert

Expired

APICS CSCP MCSA: BI Reporting MCT: Microsoft Certified Trainer

Awards & recognition

Community Super User ONGOING

Microsoft MVP ONGOING

2018

2020

2022

2024

NOW

Thesis Excellence Award

Rotterdam Business School

Hall of Fame

Maven Analytics

Community Spotlight

Power BI Community

Community

Blog & community.

Sharing knowledge through writing, mentoring, and open-source contributions.

articlepowerbimvp.com

Deep dives into Power BI, PBIR, DAX, Microsoft Fabric, and DevOps for BI. Discoveries, patterns, and techniques from real enterprise work.

Visit blog arrow_forward

volunteer_activismCommunity impact

As a Microsoft MVP and Power BI & Fabric Super User, I actively contribute through mentoring, forum support, and knowledge sharing.

2,700+ accepted solutions in the Power BI & Fabric Community
9,000+ LinkedIn followers

linkProfiles & links

Find me across the Microsoft ecosystem:

Targets & objectives

What I'm building toward in the AI era of enterprise BI.

PBIR goes GA this quarter. Fabric adoption keeps growing. And every Copilot rollout makes the same point: AI in analytics is only as good as the semantic model underneath it. Getting governance, DevOps, and semantic architecture right now is what makes AI answers dependable later.

"Most of my work is governance and performance at scale. I build the platform, the tooling, and the working habits that keep enterprise BI reliable for people and for AI agents."

My three objectives:

AI-ready semantic layers, Governed, documented, source-controlled models that Copilot and AI agents can answer from. At IKEA this means 15+ semantic models, with my open-source documentation tooling keeping them readable.
Fabric platform architecture at global enterprise scale, Direct Lake, PBIR/TMDL governance, and CI/CD as the platform standard. I run this at IKEA today: 30,000+ users, billions of rows, over 90% of refresh load removed.
Agentic BI engineering culture, AI agents in every phase of development, with the team coached to use them well. At IKEA this roughly tripled our development speed, and the whole team works this way now.

scheduleWhat I'll deliver

Enterprise PBIR migration strategy, Plan and run your PBIR transition as the format reaches GA, with the tooling and migration patterns I built for IKEA-scale deployments.
Fabric Center of Excellence, Set the governance standards, CI/CD pipelines, and semantic model patterns that hold up across your organization.
Agentic BI engineering, Bring AI agents, from Claude to the Power BI Desktop Bridge and Skills for Fabric agent skills, into your BI development workflow the way I did at IKEA, where it roughly tripled development speed.
Semantic layer as product, Turn semantic models from report-backing datasets into governed, reusable data products that people, applications, and AI agents all read from.

Get in touch

Let's discuss your Fabric strategy.

I am exploring Power BI and Microsoft Fabric architecture roles at leading global enterprises. Based in Amsterdam, with EU work rights, open to relocation and remote. If your team faces a Fabric migration, a PBIR transition, or wants semantic models that AI can answer from, let's talk.

Power BI / Fabric Architect · Fabric Platform Lead · Amsterdam · Global

mailjonathan.jihwankim@gmail.com call06-15243304 personLinkedIn codeGitHub publicBlog descriptionDownload CV articleCareer Description

The enterprise Fabric transition is happening now. Here is what I have already shipped.

I architect the platform. I also build the tooling for it.

Tools enterprises need for PBIR/TMDL governance.

Enterprise-scale impact.

What I bring to the platform.

analyticsSemantic model architecture

account_treeDevOps & governance (PBIP/TMDL)

buildDAX & performance engineering

boltAgentic BI engineering

Microsoft BI Platform

Data & DevOps

Achievements & certifications.

Certifications

Awards & recognition

Blog & community.

What I'm building toward in the AI era of enterprise BI.

Let's discuss your Fabric strategy.