Ajete IsakuData Engineer Kosovo
Building Scalable Data Platforms and Intelligent Analytics Solutions
Databricks-certified Data Engineer turning complex data into trusted business insights — open to remote roles worldwide.
- Architecture Patterns
- 10
- Technologies in Toolkit
- 25+
- Projects Completed
- 12
- Pipelines Built
- 45+
- ExtractOracle ERP · SQL Server POS
- Validate42 data quality rules
- TransformSilver → Gold · SCD Type 2
- LoadFactSales + 7 dimensions
- ServePower BI semantic model
- Rows
- 0
- DQ pass
- 99.8%
- Duration
- 1h 25m
01 — About
Data engineering, built on trust.
I came to data engineering through the questions a business asks every morning — what did we sell, what is running low, which customers are we losing — and the realisation that every answer is only as good as the pipeline behind it. Early on I spent more time reconciling numbers that disagreed across reports than analysing them, and that pulled me upstream: into source systems, SQL, data models and the pipelines that connect them.
Since October 2026 I have been a Junior Data Engineer at Aralytiks, building warehouse loads, ingestion pipelines and Power BI-ready data models on the Microsoft data platform. My day-to-day is SQL Server and Oracle SQL, Azure Data Factory orchestration, Python for automation and validation, and a growing amount of Spark on Databricks and Microsoft Fabric. I care about the unglamorous parts — idempotent loads, reconciliation checks, clear naming, documentation — because that is what makes a number trustworthy.
What drives me is the moment a stakeholder stops exporting to Excel and starts trusting the dashboard. I am deliberately building depth in lakehouse architecture, Delta Lake and dimensional modelling, with a clear goal: to become a Senior Data Engineer and, ultimately, a Data Architect who designs enterprise platforms on Fabric, Azure and Databricks.
Mission
Design scalable, reliable, and business-driven data solutions that transform raw data into actionable insights.
- Role
- Junior Data Engineer
- Based in
- Kosovo · CET · UTC+1
- Platforms
- Azure · Microsoft Fabric · Databricks
- Currently learning
- Delta Lake internals · Unity Catalog · Fabric Real-Time Intelligence
Career goal
Grow into a Senior Data Engineer and Data Architect specialising in Microsoft Fabric, the Azure Data Platform, Databricks, Big Data and Enterprise Data Architecture.
Dimensional modelling
Kimball star schemas, conformed dimensions and SCD Type 2 history that analysts can actually use.
SQL performance tuning
Execution plans, indexing, columnstore and set-based rewrites that turn minutes into seconds.
Reliable pipelines
Metadata-driven, idempotent loads with logging, alerting and reruns that do not need heroics.
Quality by design
Validation and reconciliation built into the pipeline, not bolted on after an incident.
Business translation
Turning stakeholder questions into data requirements, grains and measurable KPIs.
Relentless learning
A structured certification and lab roadmap across Azure, Fabric and Databricks.
02 — Technical skills
A toolkit for the modern data stack.
An honest self-assessment across 26 technologies — strongest in SQL, warehousing and Azure orchestration, deliberately growing in Spark, Databricks and Fabric.
Capability profile
Average proficiency per category
- Advanced80+
- Proficient70–79
- Working60–69
- Developing< 60
Programming
The languages every pipeline, model and check is written in.
- SQLAdvanced · 86
T-SQL, PL/SQL, window functions, CTEs, MERGE, query tuning
- PythonProficient · 72
pandas, PySpark, automation, API ingestion, data validation
Data Engineering
Moving, transforming and serving data at production reliability.
- ETLAdvanced · 82
Incremental loads, staging patterns, SCD handling, error logging
- ELTProficient · 76
Load-then-transform in warehouse and lakehouse engines
- Apache SparkWorking · 62
DataFrames, partitioning, joins, caching, Spark UI analysis
- PySparkWorking · 65
Transformations, Delta MERGE, Structured Streaming basics
- DatabricksWorking · 62
Notebooks, Workflows, Auto Loader, Delta Lake, Unity Catalog
- Microsoft FabricWorking · 66
Lakehouse, Warehouse, pipelines, notebooks, Direct Lake
Cloud
Building on the Azure data platform end to end.
- AzureProficient · 70
ADLS Gen2, Key Vault, managed identities, resource organisation
- Azure Data FactoryProficient · 78
Metadata-driven pipelines, ForEach, triggers, monitoring
- Azure Synapse AnalyticsDeveloping · 58
Dedicated & serverless SQL pools, distribution strategy
Databases
Relational engines and analytical warehouses.
- SQL ServerAdvanced · 84
Stored procedures, indexing, columnstore, execution plans
- Oracle SQLProficient · 78
Source extraction, PL/SQL, analytic functions
- Data WarehousingAdvanced · 80
Staging → integration → presentation layer design
- SnowflakeDeveloping · 56
Virtual warehouses, stages, COPY INTO, time travel
Analytics
Where data becomes decisions.
- Power BIAdvanced · 80
Star-schema semantic models, DAX measures, RLS, incremental refresh
Data Architecture
Patterns that keep platforms understandable as they grow.
- Kimball MethodologyAdvanced · 80
Bus matrix, grain declaration, conformed dimensions
- Star SchemaAdvanced · 84
Fact/dimension design, surrogate keys, role-playing dates
- Snowflake SchemaProficient · 75
Normalised hierarchies and when not to use them
- Delta LakeWorking · 64
ACID tables, MERGE, time travel, OPTIMIZE, clustering
- Medallion ArchitectureProficient · 72
Bronze / Silver / Gold contracts and ownership
Engineering Practices
What keeps a platform healthy after go-live.
- GitProficient · 74
Branching, pull requests, code review, versioned SQL
- Azure DevOpsWorking · 62
Repos, boards, YAML pipelines for ADF and database deploys
- Data QualityProficient · 76
Rule frameworks, reconciliation, anomaly checks, scorecards
- Data GovernanceWorking · 62
Ownership, lineage, access control, PII handling
- Performance OptimizationProficient · 74
Plans, indexing, partitioning, file sizing, cost tuning
03 — Featured projects
Enterprise-grade case studies.
Five end-to-end deliveries, each documented like an architecture review: problem, design, data model, code, trade-offs and measured impact. Client names are anonymised.
04 — Data architecture
The patterns behind reliable platforms.
Ten reference architectures I design and build with — explore each diagram interactively to see how data moves, where responsibilities sit, and when each pattern fits.
Medallion Architecture
Progressive refinement: raw → cleansed → business-ready.
Select a component for details
How it works
Data moves through bronze, silver and gold layers, each with a clear contract. Bronze preserves the raw truth for replay, silver delivers clean conformed entities, and gold serves business-shaped tables to BI and ML.
When to use: Lakehouse platforms on Databricks or Microsoft Fabric with multiple sources and consumers.
Design principles
- Bronze is append-only and replayable
- Silver owns typing, deduplication and conformance
- Gold is modelled for its consumers
- Every layer is a Delta table with ACID guarantees
Component detail
Click any stage or table in the diagram to see what it is responsible for.
05 — Experience
Delivering outcomes, not just pipelines.
What I work on day to day at Aralytiks — and the standard I hold each area to.
Junior Data Engineer
Aralytiks
Kosovo
Joined Aralytiks in October 2026 to build data warehouse, ingestion and reporting solutions on the Microsoft data platform — from source analysis and SQL development through pipeline operations and Power BI enablement.
- SQL Server
- Oracle
- Azure Data Factory
- Microsoft Fabric
- Databricks
- Python
- Power BI
- Azure DevOps
- Git
SQL Development
Write and review T-SQL and Oracle SQL: stored procedures, MERGE-based loads, window-function analytics and reporting views.
Readable, set-based SQL that passes code review
ETL Development
Build Azure Data Factory and Fabric pipelines with incremental watermarks, parameterised datasets and reusable templates.
Reusable, parameterised pipelines over one-off builds
Pipeline Monitoring
Support daily run health: failure triage, reruns, alerting rules and run logs that make root causes visible.
Failures caught before report users notice
Data Warehouse Development
Model and load Kimball star schemas — conformed dimensions, SCD Type 2 history, snapshot and transaction facts.
Grain declared and agreed before any ETL is written
Data Quality Management
Implement validation rules, source-to-target reconciliation and quality checks shared with data owners.
Reconciliation built into every load
Performance Tuning
Analyse execution plans, apply columnstore and covering indexes, rewrite row-by-row logic as set-based SQL.
Every change measured before and after
Reporting Support
Build star-schema semantic models, DAX measures and row-level security for Power BI report developers.
Models report developers can use without workarounds
Stakeholder Collaboration
Join requirement sessions and translate business questions into grains, KPIs and acceptance tests.
KPI definitions agreed and documented up front
Previous roles
Feb 2026 — Sep 2026
Junior Data Engineer
Genpact
- Built and maintained scalable pipelines on Azure Databricks, PySpark, Python and SQL for enterprise analytics and reporting.
- Implemented Medallion Architecture (Bronze, Silver, Gold) to deliver business-ready datasets.
- Owned data documentation, metadata and governance standards; configured Databricks Genie Spaces for natural-language self-service analytics.
Nov 2025 — Jan 2026
Onboarding, Data Management & BI
ProCredit Bank
- Translated reporting requirements into BI solutions; built and maintained MIS reports and dashboards for operational and strategic decisions.
- Analysed performance and financial aspects of banking products, monitoring trends and key metrics.
Jul 2025 — Oct 2025
Investment Analyst Intern (Reporting & Data Analyst)
Deloitte Switzerland
- Prepared investment and financial reports; consolidated and cleaned financial data from multiple sources.
- Tracked key metrics and presented findings to senior colleagues.
How I work
Idempotent by default
Every load can be re-run safely. Reruns are a button, not a war room.
Reconcile before you publish
Numbers are checked against source before anyone sees them.
Model the business, not the source
Grains and dimensions follow how the business thinks.
Observable pipelines
Every run logs rows, duration and watermark — so issues are found, not reported.
06 — Certifications
Certified, and still building.
Databricks-certified today, with a clear roadmap across Microsoft Azure, Fabric and Power BI.
Databricks · Professional
Databricks Certified Data Engineer Associate
Building ETL pipelines with Spark SQL and PySpark, Delta Lake, the Medallion architecture and production workflows on the Databricks platform.
- PySpark
- Delta Lake
- Medallion
- ETL
Databricks · Partner Academy
Databricks Partner Training
Lakehouse advantages, Data & AI governance, Gen AI with Data Intelligence, Financial Services and Healthcare specialisations, Spark and ETL migration.
- Lakehouse
- Governance
- Gen AI
- Migration
ICK · Course
Python
Python programming fundamentals applied to data processing and automation.
- Python
- Automation
ICK · Course
Machine Learning
Machine learning foundations and model building, complementing the MSc in Data Science & AI.
- Machine learning
- Modelling
Microsoft · AZ-900
Microsoft Azure Fundamentals
Core cloud concepts, Azure architecture, services, security, governance and cost management.
- Cloud concepts
- Azure services
- Governance
- Pricing
Microsoft · DP-600
Microsoft Fabric Analytics Engineer Associate
Planning, implementing and managing analytics solutions with Fabric lakehouses, warehouses and semantic models.
- Lakehouse
- Warehouse
- Semantic models
- Direct Lake
Microsoft · DP-700
Fabric Data Engineer Associate
Designing and implementing data ingestion, transformation, security and monitoring across Microsoft Fabric.
DP-203 retired on 31 March 2025; DP-700 is its successor.
- Data Factory
- Spark
- Real-time
- Security
Microsoft · PL-300
Power BI Data Analyst Associate
Preparing, modelling, visualising and analysing data with Power BI, and managing secure deployments.
- Data modelling
- DAX
- Power Query
- Row-level security
07 — Impact dashboard
Engineering impact, measured.
The numbers I track across my delivery work — reliability, quality, performance and adoption over the last 12 months.
Projects Delivered
12
+5 in the last 12 months
Warehouse, pipeline and reporting deliveries
ETL Pipelines Built
45+
+19 in the last 12 months
Batch and incremental pipelines in production
SQL Queries Optimized
120+
+48 in the last 12 months
Procedures, views and report queries tuned
Data Processed
1.8 TB/mo
2× vs. a year ago
Across warehouse and lakehouse loads
Performance Improvement
62%
Avg. runtime reduction
Measured before vs. after tuning
Data Quality Score
98.7%
+7.7 pts since baseline
Rules passed across monitored tables
Pipeline Reliability
99.4%
Above 99% SLA target
Successful runs, trailing 12 months
Dashboard Adoption
85%
+23 pts after model redesign
Weekly active vs. licensed report users
Operational trends · Nov 2025 – Oct 2026
Pipeline reliability
Monthly success rate vs. 99% SLA target
Show data tableHide data table
| Month | Success rate (%) |
|---|---|
| Nov | 98.1 |
| Dec | 98.4 |
| Jan | 98.2 |
| Feb | 98.9 |
| Mar | 99 |
| Apr | 99.1 |
| May | 99.3 |
| Jun | 99.2 |
| Jul | 99.5 |
| Aug | 99.4 |
| Sep | 99.6 |
| Oct | 99.4 |
Data processed
Terabytes loaded per month across all pipelines
Show data tableHide data table
| Month | Data processed (TB) |
|---|---|
| Nov | 0.9 |
| Dec | 1 |
| Jan | 1.05 |
| Feb | 1.1 |
| Mar | 1.2 |
| Apr | 1.3 |
| May | 1.35 |
| Jun | 1.45 |
| Jul | 1.5 |
| Aug | 1.6 |
| Sep | 1.7 |
| Oct | 1.8 |
Query performance tuning
P95 runtime in seconds, before vs. after optimisation
Show data tableHide data table
| Workload | Before (s) | After (s) |
|---|---|---|
| Sales reporting | 48 | 3.1 |
| Customer 360 | 64 | 9.5 |
| Ops dashboard | 35 | 4.2 |
| Inventory snapshot | 126 | 22 |
| Finance close | 210 | 41 |
Data quality by dimension
Share of rule checks passed across monitored tables
Show data tableHide data table
| Dimension | Score (%) |
|---|---|
| Uniqueness | 99.8 |
| Completeness | 99.2 |
| Consistency | 98.5 |
| Validity | 98.4 |
| Timeliness | 97.6 |
08 — Writing
Notes from the data engineering trenches.
Practical articles on modelling, pipelines, platforms and performance — what worked, what didn't, and why.
09 — Contact
Let's build data platforms people trust.
Recruiters and hiring managers — I'm looking for a team where I can ship reliable data products on Azure, Microsoft Fabric and Databricks, and grow towards data architecture. I'd love to hear about the problems your data team is solving.
Target roles
- Data Engineer
- Databricks / Lakehouse Engineer
- Analytics Engineer
- BI Engineer
- Remote (worldwide)
- Hybrid
- Open to relocation
- Kosovo
- CET · UTC+1
- Replies within 24 hours