raw_events · unformatted, unfiltered, unmistakably me · Colombo, Sri Lanka

Engineering data for better decisions.

I'm Ruwan Pathirana, a cloud data engineer with experience designing, supporting, and optimizing analytics solutions on Microsoft Fabric and Azure. I enjoy solving data challenges and building scalable pipelines that deliver trusted insights.

See the work Get in touch
PROFILEprofiled 1 record

I started in data on the reporting side building dashboards that business teams actually depend on and gradually moved upstream as I became more involved in understanding and fixing data issues closer to the source.

Today I work as an Associate Consultant in Data & AI, where I split my time between developing Power BI solutions and supporting Azure-based data platforms for enterprise clients across multiple regions. The role combines building, maintaining, and troubleshooting production data systems to keep analytics reliable and trusted.

Outside of client work, I explore machine learning and generative AI to deepen my understanding of how end-to-end data and AI systems behave in real scenarios.

df.describe()n=1
current_roleAssociate Consultant, Data & AI
educationB.Sc(Hons) IT & Mgmt
institutionUniversity of Colombo
certifications3X Microsoft Certified
core_stackAzure · Fabric · Power BI
TRANSFORM3 tasks · DAG order
Associate Consultant - Data & AI
Fortude
Dec 2024 — Present
RUNNING
  • Design, build, and maintain Power BI data models and dashboards that support business reporting and day-to-day decision-making.
  • Provide managed support for Azure-based cloud data platforms, troubleshooting and resolving issues for multi-region enterprise clients.
  • Support analytics solutions built on Microsoft Fabric, keeping data delivery and pipeline performance reliable.
  • Maintain an Infor BI ERP solution — monitoring, troubleshooting, and front-line customer support for enterprise users.
  • Work directly with clients to understand issues, investigate root causes, and ship fixes that hold.
upstream: Intern Data Scientist @ Epic Lankafeeds_into: current
Trainee - Data Scientist
Epic Lanka Technology
Nov 2023 — Nov 2024 · 1yr
SUCCESS
  • Supported development of cloud-based ETL pipelines using Azure Data Factory, Databricks, and Synapse.
  • Built and maintained dashboards in Power BI, Tableau, and D3.js for financial analytics.
  • Implemented document classification and OCR automation.
  • Assisted in developing machine learning models for risk assessment and prediction, including API deployment.
  • Contributed to chatbot development for business use cases.
upstream: Trainee @ Bank of Ceylonfeeds_into: Associate Consultant @ Fortude
Trainee
Bank of Ceylon
Feb 2019 — July 2019 · 6mo
SUCCESS
  • Supported development of cloud-based ETL pipelines using Azure Data Factory, Databricks, and Synapse.
  • Built and maintained dashboards in Power BI, Tableau, and D3.js for financial analytics.
  • Implemented document classification and OCR automation.
  • Assisted in developing machine learning models for risk assessment and prediction, including API deployment.
  • Contributed to chatbot development for business use cases.
upstream: none - source nodefeeds_into: Trainee Data Scientist @ Epic Lanka
CURATED4 tables · self-directed
gold.ms_fabric_end_to_end
end-to-end
problem
Wanted real, production-shaped experience with Microsoft's unified analytics platform — not just the parts covered on the certification exams.
approach
Ingested data from an external API straight into OneLake, transformed it with PySpark, then built the reporting layer on top in Power BI.
stack
Microsoft Fabric, OneLake, PySpark, Power BI
status
Complete — full ingest-to-dashboard path, self-directed.
gold.fraud_gan_balancer
prototype
problem
Credit card fraud datasets are heavily imbalanced, which quietly sabotages model performance on the cases that actually matter.
approach
Built a Generative Adversarial Network to synthesize realistic fraudulent transactions, balancing the training set with real and generated samples.
stack
Python, TensorFlow / Keras, GANs
status
Working prototype, portfolio project.
gold.sem_image_classifier
prototype
problem
Scanning Electron Microscopy images need classification, but off-the-shelf models don't generalize well to specialized lab imagery.
approach
Built a feature-driven pipeline: extracted a diverse set of features from labeled SEM images, then trained classical ML models on top of them.
stack
Python, Scikit-learn, feature engineering
status
Working prototype, portfolio project.
gold.voice_emotion_shift
research
problem
Most emotion recognition treats a clip as one static label — but emotion shifts mid-conversation, and that transition itself carries signal.
approach
Applying deep learning and clustering techniques to identify where and how one emotion transitions into another within continuous audio segments.
stack
Python, deep learning, clustering
status
Ongoing research.
VALIDATE5 / 5 checks passed
✓
Power BI Data Analyst Associate
PL-300 · Microsoft Certified
PASSED
✓
Fabric Data Engineer Associate
DP-700 · Microsoft Certified
PASSED
✓
Fabric Analytics Engineer Associate
DP-600 · Microsoft Certified
PASSED
✓
Supervised Machine Learning
DeepLearning.AI
PASSED
✓
Machine Learning with Python
LinkedIn Learning
PASSED
ORCHESTRATE7 layers
LANGUAGES & QUERYING
PythonSQLR
CLOUD & ANALYTICS
AzureMicrosoft FabricDatabricksPySparkSnowflakeInfor BISSAS
INTEGRATION
Azure Data FactoryInfor DWDFivetran
DATABASES
PostgreSQLMySQLMS SQLMongoDB
VISUALIZATION
Power BITableauInfor Office PlusExcelJasper Studio
AI & ML
TensorFlowScikit-learnKerasOCR
TOOLING
GitHubBitbucketAzure DevOpsJIRASysAidFreshDesk
SERVE200 OK

This is the layer everything upstream exists to support. Open role, a Fabric or Power BI project, or just want to talk about GANs over coffee — send a request.

GET /contact
200 OK · 12ms
{
"email": rpathirana836@gmail.com,
"linkedin": in/ruwanpathirana,
"github": github/ruwanpathirana,
"location": Colombo, Sri Lanka
}