MALINDU GIMHANA PORTFOLIO ’26

Data, with
dimension.

I turn complex data
into intelligent systems.

Data Science undergraduate. Engineer by mindset.
Visual thinker by background.

Explore my work Download CV

+ COMPUTATIONAL STUDY — 001INTERACTIVE STUDY

XYSRI LANKA
UTC +05:30

RELATIONSHIPS EMERGE

Find the relationships. Reveal the shape.

01A DIFFERENT PERSPECTIVE

Data tells a story.
Engineering makes it useful.
Design makes it understood.

I’m Malindu, a third-year Data Science undergraduate at SLIIT. I work across statistical analysis, machine learning, data engineering and software development.

Before that, I built worlds in 3D. Working in visualization taught me to see relationships, care about detail and make complex ideas tangible. That perspective still shapes everything I build.

A little more about my path

01 / ANALYZE

02 / ENGINEER

03 / COMMUNICATEONE CONNECTED PRACTICE

02SELECTED WORK

Questions explored.
Systems built.

THREE PROJECTS. DIFFERENT LENSES.

From statistical evidence to production data.
A closer look at how I think and build.

01/ 03STATISTICAL CONSULTANCYR / PYTHON

The human side of data.

Workplace Stress Factor Analysis

22,750 employee records. One important question: how can we recognize workplace stress before it becomes burnout?

FIG. 01 / UNDERSTANDING THE INDIRECT EFFECT↗

WorkloadPsychological
mediatorBurnout

Expected misclassification costINDEX / BASELINE = 100

Naive baseline

100

Cost-sensitive model

CROSS-VERIFIED IN R + PYTHONREPORTED PROJECT RESULTS

22,750EMPLOYEE RECORDS

67.6%MEDIATED EFFECT

38%LOWER EXPECTED COST

19PEER-REVIEWED SOURCES

THE TAKEAWAY

Good analysis goes beyond significance. It helps people make better decisions.

02/ 03DATA ENGINEERING & BUSINESS INTELLIGENCESQL / POWER BI

From scattered data
to a single picture.

Brazilian E-Commerce Data Warehouse

A complete journey from messy source files to a dimensional warehouse and business intelligence that answers real questions.

FIG. 02 / THE PATH FROM DATA TO DECISIONS↗

04 / WAREHOUSE

Star-schema architecture with surrogate keys and a role-playing date dimension.

ROW-COUNT + REFERENTIAL-INTEGRITY CHECKSSELECT A STAGE TO EXPLORE

1M+ROWS PROCESSED

~100KCUSTOMER ORDERS

5 + 2DIMENSIONS + FACT TABLES

7VALIDATED BUSINESS INSIGHTS

THE TAKEAWAY

A useful dashboard starts long before the first chart. It starts with trustworthy data.

03/ 03APPLIED MACHINE LEARNINGMAS HOLDINGS

Intelligence on
the factory floor.

Smart BOM Predictor · MAS Holdings — Silueta

Predicting jig manufacturing costs and material consumption from historical BOM and production data. Built with the people who use it.

FIG. 03 / MANUFACTURING INTELLIGENCE↗

INPUT

Historical BOM

Production data

Smart BOMPREDICTION ENGINE

OUTPUT

Manufacturing cost

Material consumption

PREPROCESSINGFEATURE ENGINEERINGMODEL DEVELOPMENTREPORTING

BUILT WITH PROCESS ENGINEERSMAS HOLDINGS — SILUETA

BUILT IN A REAL MANUFACTURING ENVIRONMENT

Data preprocessing. Feature engineering. Model development.
And the domain knowledge that only comes from working with process engineers.

Further explorations / 02

MORE WAYS TO MAKE DATA USEFUL

MACHINE LEARNING · BACKEND

Predictive Maintenance Engine

Random Forest / Flask / REST API

A Random Forest model built from historical fleet data, served through a Flask REST API and integrated with a MEAN-stack logistics platform for real-time maintenance predictions.

The workflow includes data preprocessing, feature engineering and model evaluation, followed by API integration into the logistics platform.

NATURAL LANGUAGE PROCESSING

Sentiment Analysis ML Model

XGBoost / TF-IDF / NLP

From unstructured language to sentiment predictions: a text classification pipeline built with TF-IDF and XGBoost.

The project combines NLP text preprocessing, TF-IDF feature extraction, XGBoost classification, hyperparameter tuning and model evaluation.

03THE PATH SO FAR

A visual beginning.
An analytical direction.

Different environments.
The same instinct to understand how things work.

FEB 2026 — AUG 2026 07 MONTHS

MAS Holdings — Silueta

Industrial Engineering Intern

Applied data science in a live manufacturing environment. Developed the Smart BOM Predictor to estimate jig manufacturing costs and material consumption, working closely with process engineers.

Applied MLManufacturing analyticsProduction planning

See the work

SEP 2023 — JUN 2024 10 MONTHS

VAEG

Visual Art and Experiences Group

Assistant 3D Generalist

Created professional 3D visualizations, event assets and interactive experiences using Unreal Engine. Worked across client projects, production workflows and multidisciplinary teams.

3D visualizationUnreal EngineInteractive experiences

A PERSPECTIVE THAT STUCK

Make the complicated feel intuitive. Make the details count.

04THE TOOLKIT

A connected
way of thinking.

The tools change. The foundations stay:
reason clearly, build carefully, keep learning.

PYTHONAcross statistical analysis, data engineering & machine learningSELECT A TOOL TO FIND THE CONNECTION

01Programming

02Machine learning

03Statistical analysis

04Data warehousing & BI

05Backend

06Databases

07Development

05FOUNDATIONS & CURIOSITY

FORMAL EDUCATION

Learning the theory.
Building the practice.

BSc (Hons) Information Technology

Specializing in Data Science

Sri Lanka Institute of Information Technology — SLIIT

3.49GPA / 4.0

THIRD YEAR

Data Analytics · Statistics · Data Warehousing · Database Management · OOP · Software Engineering

CONTINUED EXPLORATION

NEO4J / 2024

Certified
Professional

Neo4j certification

NEO4J / 2024

Graph Data
Science

Neo4j certification

06WHAT COMES NEXT

BASED IN SRI LANKA · OPEN TO OPPORTUNITIES

Good questions.
Better possibilities.

Have a difficult problem or an interesting opportunity?
Let’s turn it into something useful.

hello.malindu@gmail.com

Data, withdimension.

I turn complex datainto intelligent systems.

Data tells a story.Engineering makes it useful.Design makes it understood.

Questions explored.Systems built.