Hello! 👋

I'm Sakuna Harinda Jayasundara

Software Engineer & Researcher

About Me

Ph.D.-qualified software Engineer and Researcher with 5+ years of industry and research experience building scalable, production-grade systems across backend development, machine learning, and automation. Brings 8+ years of experience in Python and Java, along with extensive work across modern ML frameworks, containerization, orchestration, CI/CD pipelines, and end-to-end test automation. Ph.D. from the University of Auckland, with research spanning deep learning, natural language processing, cybersecurity, and HCI. Focused on applying data-driven and research-informed approaches to solve real-world engineering problems.

Pytorch Tensorflow Keras Python Java C/C++ Springboot Docker Kubernetes

Experience

Software Engineer

H2O.ai

June 2022 - October 2022
  • Built and maintained scalable Wave applications to support internal and external ML workflows, enhancing reliability and developer productivity through automation.
  • Designed and implemented robust end-to-end test automation pipelines using Python, Jenkins, and Groovy, reducing manual testing effort and increasing release velocity.
  • Led the architecture and development of a best-in-class test automation framework, including guiding implementation, defining coding standards, and shaping the automation roadmap for cross-team use.
  • Collaborated with engineering teams to integrate automation into CI/CD workflows, improving code quality and shortening time-to-deployment for key features.

DevOps Engineer

Axiata Digital Labs

July 2021 - June 2022
  • Designed, developed, and maintained backend RESTful and SOAP APIs using Java and Spring Boot to support customer-facing services for clients associated with Celcom Malaysia, ensuring high reliability and seamless integration with existing telecommunications systems.
  • Built an anomaly detection system leveraging OpenTelemetry data and machine learning models to identify performance and reliability issues in production services, enabling proactive issue detection and improved system observability.
  • Developed an NLP-based repository classifier to automatically identify public GitHub projects with potential relevance to internal initiatives, streamlining competitive analysis and reducing manual review workload.
  • Collaborated with cross-functional product and engineering teams to define API specifications, implement service integration, and ship scalable backend functionality under CI/CD workflows.

Intern Electronics Engineer

Paraqum Technologies

May 2019 - December 2019
  • Developed a GTP (GPRS Tunneling Protocol) packet analysis tool along with a test environment, enabling detailed inspection and debugging of mobile core network traffic for protocol validation and performance analysis.
  • Designed and implemented a load-balancing software module to monitor network interfaces and manage data traffic efficiently, improving throughput handling in simulated network scenarios.
  • Enhanced performance and extended functionality of the company’s AD client software, optimizing existing C++ modules for responsiveness and reliability under real-world conditions.

Education

Ph.D. in Computer Science

University of Auckland

2022 October - 2026 January
  • Introduced a human-centric approach to automated access control policy generation, addressing fundamental limitations of both manual and fully automated methods by reducing human error and AI-induced inaccuracies.
  • Developed language model-driven techniques to translate natural-language access requirements into enforceable access control policies, enabling accurate handling of complex, ambiguous, real-world specifications.
  • Designed mechanisms that support automatic error detection and recovery, enabling safe deployment of AI-generated security policies in high-risk environments.
  • Conducted empirical user studies with professional policy implementers, demonstrating improved accuracy, efficiency, trust, and adoption potential compared to existing approaches.
  • Produced comprehensive analyses and datasets that address long-standing research gaps in access control policy generation.

B.Sc. (Hons.) in Electronics and Telecommunication Engineering

University of Moratuwa

November 2026 - August 2021
  • Attained a GPA of 3.92/4.2, obtaining a First Class.
  • Dean’s List Honoree, recognized in all four academic years of the B.Sc. program.

Publications

Journal

Jayasundara, S. H., Gamagedara Arachchilage, N. A., & Russello, G. (2024). SoK: Access control policy generation from high-level natural language requirements. ACM Computing Surveys, 57(4), 1-37.

Conference

Jayasundara, S. H., Arachchilage, N. A. G., & Russello, G. (2024). "AccessFormer": Feedback-driven access control policy generation framework. In the Symposium on Usable Security and Privacy (USEC).

Journal

Jayasundara, S. H., Gamagedara Arachchilage, N. A., & Russello, G. (2026). AGentVLM: Access Control Policy Generation and Verification Framework with Language Models. Journal of Information Security and Applications.

Conference

Abeywardena, K., Sumanthiran, S., Jayasundara, S. H., Karunasena, S., Rodrigo, R., & Jayasekara, P. (2023, September). KORSAL: Key-Point Based Online Real-Time Spatio-Temporal Action Localization. In 2023 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE) (pp. 279-284). IEEE.

In Submission

Jayasundara, S. H., Gamagedara Arachchilage, N. A., Biddle, R., & Russello, G. (2026). CHAGent: Context-aware Human-centric Access Control Policy Generation. IEEE Symposium on Security and Privacy.

In Submission

Jayasundara, S. H., Gamagedara Arachchilage, N. A., Biddle, R., & Russello, G. (2026). "This is going to change the game": Design and Evaluation of a Usable and Accurate Automated Access Control Policy Generation System. ACM CHI conference on Human Factors in Computing Systems.

Projects

Ragatouille

The Ultimate Guide to Mastering Retrieval-Augmented Generation (RAG) with LangChain. Featured in LangChain LinkedIn page!.

NLPPythonLangchainNeo4J