Sakuna Harinda
Jayasundara

I build scalable, production-grade systems across backend, machine learning, and automation — 5+ years of industry and research experience, and a Ph.D. from the University of Auckland with research spanning deep learning, natural language processing, cybersecurity, agentic security, & HCI.

Engineer and researcher who turns research-grade ideas into production systems.

I have 5+ years of combined industry and research experience building scalable, production-grade systems across backend development, machine learning, and automation. My toolbox spans 8+ years of Python, Java, C/C++, TypeScript, and C#, plus deep work in agentic AI, agentic security, modern ML/MLOps frameworks, CI/CD pipelines, and end-to-end test automation.

My Ph.D. at the University of Auckland spans deep learning, natural language processing, cybersecurity, agentic security, and HCI. I'm driven by applying data-driven, research-informed approaches to real-world engineering problems.

• ▹Prepare and deliver lectures across Data Structures & Algorithms, Software Engineering Theory, Cryptographic Management, and Programming for Industry (Java).
• ▹Supervise examinations and assess assignments and projects, ensuring fair, consistent evaluation — alongside my Ph.D. research.

• ▹Built and maintained scalable web applications supporting internal and external ML workflows, boosting reliability and developer productivity through automation.
• ▹Designed robust end-to-end test automation pipelines with Python, Jenkins, and Groovy — cutting manual testing effort and increasing release velocity.
• ▹Led architecture of a best-in-class test automation framework, defining coding standards and the automation roadmap for cross-team use.

• ▹Designed and maintained backend RESTful and SOAP APIs in Java & Spring Boot for customer-facing services (Celcom Malaysia), with high reliability and seamless telecom integration.
• ▹Built an anomaly detection system on OpenTelemetry data and ML models to surface production performance issues proactively.
• ▹Developed an NLP-based repository classifier to flag relevant public GitHub projects, streamlining competitive analysis.

• ▹Built a GTP (GPRS Tunnelling Protocol) packet analysis tool and test environment for inspecting and debugging mobile core network traffic.
• ▹Implemented a load-balancing module to monitor interfaces and manage traffic, and optimised C++ modules for responsiveness and reliability.

• ▹Introduced a human-centric approach to automated access control policy generation, reducing both human error and AI-induced inaccuracies.
• ▹Designed LLM-driven workflows transforming natural-language requirements into structured, validated outputs via context retrieval, generation, validation, and user feedback.
• ▹Built ML-based error detection and recovery for safe deployment of AI-generated security policies in high-risk environments.
• ▹Resulted in 3 publications (one flagship cybersecurity conference, two Q1 journals) with 2 more under review at flagship CS & HCI venues.

• ▹GPA 3.92/4.2 — First Class; Dean's List Honouree across all four academic years.
• ▹Built a deep-learning maritime surveillance system for the Sri Lanka Navy (PyTorch, PyQt) — 2nd runner-up, National ICT Awards for Best University Project.

Selected publications

AgenticSLR

A web-based, agentic systematic literature review platform built on LangGraph — reducing the manual burden every first-year Ph.D. student faces.

Maritime Surveillance

Deep-learning system for the Sri Lanka Navy detecting suspicious maritime activity. 2nd runner-up, National ICT Awards.

Let's build something.

I'm open to Software Engineer and ML / AI Engineer opportunities. Whether it's a role, a collaboration, or a question — my inbox is always open.