R. LUMINDA KULASIRI
Kennesaw GA
Cell: 770-235-6974
| email:
Status: US Citizen
Github: https://github.com/lumindak
Linkedin:https://www.linkedin.com/in/r-luminda-kulasiri-61b8551a2/
DATA SCIENTIST/PHYSICIST
Results-oriented Data Scientist and Physicist with a PhD in Physics and over ten years of experience in academia and industry. Specialized in building machine learning models and proficient in data analytics.
TECHNICAL SKILLS
· Programming: Python, JavaScript, R, C++, C, Java
· Statistical data Analysis : R, Python, Excel, VBScript, ROOT
· Cloud Platforms: AWS, sagemaker, Google Colab
· Big Data: MapReuce, Hive, Hadoop, Spark
· Workflows: Alteryx
· Natural Language Processing (NLP)
· Databases: SQL, MongodB
· BI tools : Tableau, Java Script
· Shell Machine Learning models : Sklearn, TensorFlow, PyTorch
· Simulation and Optimizations: Simpy, Pulp, C++
· Version Conrol: Github, Gitlab, Jira, Confluence
· Project Management: Jira, Confluence
· Data Visualization : Matplotlib, Seaborn, Plotly, ROOT, Excel
· OS: Linux, Unix, MacOS, Windows
· Teaching : Physics, Mathematics, Data Science
RECENT PROJECTS
· Early Attrition Prediction: Predict customers who stop using the credit cards at Fleetcor during the first ten weeks. Reducing the early attrition is an important part of increasing the revenue at Fleetcor. The impact of the model is estimated to be 1.01 million dollars.
· Adobe Data Feed: Monitoring, Maintenance, improving and fixing the problems when they occur. This workflow feeds the data to Adobe Analytics dashboard which is the primary tool used for digital marketing campaigns at Fleetcor.
· Customer Revenue Prediction for Adobe Analytics : This model predicts the annual revenue for each new customer. This information is used by the Fleetcor's digital marketing team to target high revenue customers. The model predictions will help increase the revenue by 10-30%.
· Revenue Prediction (Time Series Forecaasting) : I have successfully created machine learning models to predict revenue of different lines of businesses (LOBs) at Fleetcor Technology Inc.(Used ARIMA, CNN, LSTM models with Python, SQL, Alteryx). These models helped to reduced prediction error by ~10%
· Fraud Analysis : Exploratory data analysis for fraud indicators. This analysis provided valuable indicators to identify fraud.
· Lead Scoring Model: prioritize customer leads using Adobe Click Stream data at Fleetcor
· Creating a Dashboards: Display “Loss Rates” at Fleetcor Technology Inc. and data quality monitoring
· Data migration from Old Salesforce instance a to Salesforce Gold instance
· Credit Card Fraud Detection Built a model to predict credit card fraud. The model can predict fraud transactions with a 93% accuracy (Used Colab, Amazon RDS, PySpark, sklearn, TensorFlow, Keras and PostgreSQL).
· Analyzing Atlanta Crime data we were able to identify safe/unsafe neighborhoods in the city. Also, we uncovered the times and days crimes occur more frequently (Used Python, Pandas, Gmaps, JSON, Matplotlib and Scipy).
· Analyzing the Voting records: Vpting records of Georgia’s members of Congress – We were able to uncover the priorities of each member and their voting trends (Used API’s, SQLite, Flask, JavaScript, D3.js, Plotly, Canvas.js, Leaflet, Python, JSON, HTML, CSS and Bootstrap).
· Other projects: Amazon Reviews, NYK bike Data analytics, GA Members of Congress Voting records, Obesity and per Capita Income, Belly Button biodiversity (https://lumindak.github.io/projects.html)
RELAVANT EXPERIENCE
Data Scientist/Analyst 2020 - 2024
Fleetcor Technology Inc.
Built ML models, Co-owned of Adobe Analytics Workflow, Developed and maintained Adobe Analytics Workflow, Created Reports, dashboards and visualizations, Analyzed data for valuable insights
Assistant Professor 2013 - 2020
Kennesaw State University/Southern Polytechnic State University
Teaching: Introductory level and upper level physics classes
Research: Analyzing big Belle/Belle II data sets for new particle decays, simulating wind turbines, building cosmic ray detectors and analyzing data, programming with C++ and Python, data fitting, regression, visualization, neural networks, Improving the Belle II event display module
Lecturer 2008 – 2013
Southern Polytechnic State University
Taught introductory level and upper level physics classes
Adjunct Instructor 2007 -2008
Southern Polytechnic State University
Taught introductory level and upper level physics classes
Intern 2006 – 2007
Infrastructure Management Institute (IMI), Northern Kentucky University
ABAP programming
Intern 2005 – 2006
University of Cincinnati, Children’s Hospital Imaging Center
Developed a software to identify bone density using x-ray images, image processing, IDL programming, MATLAB programming
EDUCATION
2020 Georgia Tech Data Science and Analytics Boot Camp, Georgia Institute of Technology (BS equivalent), Atlanta GA
2020 Complete Data Science Course 2020, Udemy
2005 PhD, Experimental Particle Physics, University of Cincinnati, Cincinnati, OH
1996 BSc (Hons. With Mathematics minor), University of Colombo, Colombo, Sri Lanka