Developing the dockerized payment service and integrating with Stripe and PayPal and ensuring asynchronous messaging to manage payment processing workflows and ensure reliable communication between services complete with comprehensive unit tests.Designed and implemented CI/CD pipelines using Jenkins to automate build, test, and deployment processes, integrated Docker for AWS ECR, automated deployments to Amazon EKS, and managed credentials with AWS Secrets manager.

Integrating the 'group by' clause into the Ballerina language by crafting efficient syntax tree node structures and parser error handling sequences, along with the integration of essential semantic and type checking logic up to the desugaring phase with comprehensive parser and compiler testing enabling programmers to organize input data using 'grouping-keys' within Ballerina's 'query expressions'.Devoting considerable effort to align XML behavior with the Ballerina language specification, exemplified by the introduction of the 'XML sequence' concept which involved refining parser syntax trees, error handling sequences, semantic logic, and desugaring, as well as leveraging the ASM library for BIR generation, facilitating the resolution of inconsistencies in 'XML subtype' behavior within the language. Successfully addressing a range of such XML-related issues led to assuming ownership of the XML domain within the Ballerina ecosystem and providing active support to colleagues facing XML-related challenges.Rectifying discrepancies in integer behaviour by strategic repositioning of specific logic into better suited compiler phases and resolving challenges concerning union type checking of integer subtypes which consequently resolved a substantial number of integer-related issues within Ballerina ensuring overall smoothness of integer operations.

Mentoring under SCoRe Lab for the TensorMap project.

Mentoring the TensorMap project under SCoRe Lab for children aged 13 to 18.

Developing an automated code generation system for TensorFlow implementations of RNN and ML algorithms in Python, enhancing its functionality by integrating preprocessing options and generating the corresponding code and designing a user-friendly dashboard for ‘Tensormap’, enabling users to visually construct machine learning algorithms, thus empowering users to reverse engineer a visual layout into a TensorFlow implementation with minimal background knowledge.

Explored capabilities of different RNN networks within streaming contexts and conducted performance analysis using various machine learning algorithms which helped to contribute to several academic papers.Publications: Jayasinghe, M., Helmini, S., Jihan, N., Hewa, k., Perera, S. (2020). “A Machine Learning Based Approach for Predicting the Performance of Highly-Concurrent Server Applications.” International Conference on High-Performance Computing & Simulation.Helmini, S., Jihan, N., Jayasinghe, M., Perera, S. (2019). “Sales forecasting using multivariate long short term memory network models.” PeerJ Preprints 7:e27712v1, DOI:10.7287/peerj.preprints.27712v1

Worked as a front-end developer in phase 5 prepaid release of the Dialog Self Care app