Thamali Wijewardhana

Experience

• 

  WSO2 - Machine Learner

  Oct 2015 - Apr 2016

  Software Engineering Intern

  As a software engineering intern at WSO2, I have worked with various projects and had a good experience in data sciences and analytics. WSO2 is an open source middle ware company with its revolutionary componentized design and which is based on service oriented architecture. I have contributed with following areas during my internship.1. Building analytic solutions in NFL super Bowl Prediction project to predict the winner of NFL super bowl 50. NFL super Bowl is a famous foot ball match in America. In this project we predicted the winners of NFL super bowl playoff games and the final game using WSO2 machine learning concepts. This basically included researching about the game, selecting season for prediction, collecting data, decide the correct feature set(Using feature selection methods), selecting the algorithm, building and implementing the algorithm, building models, tuning for hyper parameters, predicting the winners and displaying the results in a web page.http://wso2.com/landing/big-data-game/2. Adding the Recurrent Neutral Network to WSO2 machine LearnerThis project was to add RNN to WSO2 machine learner using deeplearning4j library and spark pipeline concept.Technologies Used:Deep learning, Recurrent Neural Networks, Sentiment analysis, Word embedding, deeplearning4j library, Keras library, Java Flight Recorder (JFR), Java Mission Control (JMC), Java, Python and Apache Spark pipelines.3.Cleaning up ml REST API to make the WSO2 Machine Learner API, A RESTful APIThis project was to develop a REST based API for WSO2 machine Learner. This project basically had three steps 1.Research about REST API concepts and make the WSO2 Machine Learner API RESTful.2.Generate Swagger definition for the created REST API.3.POC on automatically generating client code on by using Swagger definition and Swagger CodegenTechnologies Used: , Apache CXF, Swagger UI, Swagger Codegen, OAuth2, Maven, Intellij and API resource selection Show less

• 

  Synopsys Inc

  Apr 2017 - now

  Currently I am working as a key member of a USA based machine learning (ML) and data analytics expert team to build innovative ML solutions. My key responsibilities include:• Surveying literature, learning, and contributing to implementing state-of-the-art techniques and tools using artificial intelligence (AI), deep learning, and ML to solve challenging problems.• Conducting experiments using generative AI to improve usage and performance of internal products.• Providing guidance, training, and technical leadership to ensure high-quality deliverables.• Designing and implementing ML pipelines to ensure efficient data processing and transformation. Show less I was mainly responsible for research and development of the below research project.Intelligent Test Selection (ITS) Research Project:ITS is a ML based solution (ensemble model) to select a subset of most probably failing tests from a regression testcase suite to resolve the high resource and time utilization happening when running the full regression. • Contributed to data preprocessing, feature engineering, data sampling, model training and tuning, automatic model tuning, model validation, and prediction flow.• Carried out machine learning experiments with researching to improve the performance of ML model.• Contributed to developing the machine learning pipeline for the product.• Designed and developed the product’s features to optimize performance, reliability, and efficiency.• Used Python, Scikit-learn, imbalanced-learn, H2O, Pandas, NumPy, Docker, Swarm, Airflow, Django, Django Rest Framework, PostgreSQL, Redis, Prometheus, React, Sass Show less Test Knowledge Database:Test Knowledge Database (TKDB) is an end-to-end solution that extracts data from devices in designverification flow, stores them after preprocessing, facilitates big data search capabilities and presentsresults to the end user. TKDB provides solutions for all four logical layers (big data sources, datamessaging and storage layer, analysis layer and the consumption layer) of a big data solution. This isan information retrieval system where users can input their queries using a custom Boolean querylanguage.• Did research and proposed improvements to the existing TKDB architecture to improve queryperformance.• Contributed to data collecting and processing python plugins, storing data in MongoDB, indexing datato Elasticsearch, querying using Elasticsearch and ES-Hadoop connector, implementing API, Kibanadashboard and the UI.• Used Java, Python, Elasticsearch, Kibana, Elasticsearch-hadoop, MapReduce, Spring, MongoDB,React, MavenData Visualization Dashboard (Full Stack):Developed a full stack system for displaying statistics of products.Languages: Javascript, PythonTechnologies: React Js, Bootstrap, D3.js, SVG, Axios, Django, Django REST FrameworkTools: PerforceAssisted intern at industrial training period and guided for success.Actively participated in code reviews (clean, high-quality, high-performance, maintainable code) Show less

  • Staff Research Engineer - Machine Learning

    Jul 2023 - now

  • Senior Research Engineer - Machine Learning

    Feb 2020 - Jul 2023

  • Research And Development Engineer

    Apr 2017 - Jan 2020