CV

Associate Technical Lead (AI R&D)

WSO2

Technical lead for AI capabilities in the WSO2 Integrator platform.

• Designing multi-agent systems and developing/training custom models for agentic workflows.
• Owning architecture, execution, and delivery of agentic developer workflows.
• Driving roadmap, coordinating contributors, and ensuring production readiness (security, scalability, reliability).
• Mentoring engineers and interns to ship high-quality AI features.

Senior Software Engineer (AI R&D)

WSO2

Led the end-to-end initiative to modernize WSO2 Micro Integrator with AI.

• MI Copilot: Owned from ideation to production. NL to Synapse DSL code generation with version-aware retrieval, XSD/Schematron validation, LSP diagnostics, and automated repair loops. 300+ active daily users.
• MI Copilot Agent Mode: Cursor/Claude Code-style agentic integration engineer inside VS Code with 22 tools across 8 categories, multi-step planning, and autonomous task execution.
• Integrated domain-specific embedding models into MI Copilot retrieval system for low-latency semantic code search.
• Optimized operational cost via advanced prompt caching (~90% cost reduction) and production hardening/observability hooks.
• Small Language Model initiative: Initiated domain-specialized SLM training for Ballerina to enable low-latency on-prem code generation.

Software Engineer

WSO2

Built open-source products for API Management and Integration.

• Delivered multiple end-to-end features for WSO2 API Manager and contributed 40+ product improvements.
• Distributed Transaction Counter: Scalable transaction counting for high-throughput API gateways handling 10,000+ TPS with sub-millisecond latency.
• GenAI / Multi-agent framework: Led a low-code framework supporting multiple LLM backends + vector stores, reducing delivery time and standardizing deployments.

Lead Research Collaborator

University of Moratuwa

Supervisors: Dr. Surangika Ranathunga (Massey University, NZ) and Dr. Nisansa de Silva (UOM)

• Leading Google-funded research on Automatic Post-Editing (APE) for low-resource languages (Sinhala, Tamil).
• Proposing controllable minimal-edit post-editing with hidden-state anchoring and dynamic edit-distance constrained decoding.
• Driving experimentation, evaluation, and research direction for practical APE workflows.

Research Engineer

Persistent Systems

• Developed architectures and POCs for a new API Marketplace intended to host and manage thousands of APIs.

Graduate Researcher (Thesis Continuation)

University of Moratuwa

Supervisors: Prof. Sanath Jayasena, Dr. Miyuru Dayarathna

• Extended undergraduate thesis on privacy-preserving distributed graph ML.
• Developed novel common-attribute and common-entity aggregation mechanisms for heterogeneous multi-organization graph learning while maintaining privacy.
• Implemented prototype on JasmineGraph distributed graph database.

Undergraduate Thesis Researcher

University of Moratuwa

Supervisors: Prof. Sanath Jayasena, Dr. Miyuru Dayarathna

• Researched privacy-preserving distributed graph ML and federated GNNs for large-scale link prediction.
• Built memory-efficient distributed/federated GCN training enabling training on very large graphs on commodity hardware (JasmineGraph).
• Peer-reviewed publication at IEEE Big Data 2020.

Machine Learning Research Intern

Zeptolytics

• Developed end-to-end automated data analytics and Business Intelligence pipeline for the ZEPTO platform.
• Resulted in 40% improvement in decision-making speed for client companies.
• Built using Python, Scikit-learn, pandas, NumPy. Deployed with Docker and AWS.

B.Sc. Engineering (Honours)

University of Moratuwa

Computer Science and Engineering

GPA: 3.89 - First Class Honours

• Dean’s List: 7 out of 8 semesters (Semesters 1, 3, 4, 5, 6, 7, 8)
• Thesis: Privacy-Preserving Machine Learning on Distributed Graphs (Grade: A+)
• University Award for Best R&D Project (highest marks achieved)

Memory-Efficient Graph Convolutional Network-Based Distributed Link Prediction

Anomaly Detection using Deep Reconstruction and Forecasting for Autonomous Systems

IEEE Signal Processing Cup 2020 - First Runner-Up

• First Runner-Up at ICASSP 2020 finals (team Icarus Inhibition).
• Achieved the highest score in the selection round for anomaly detection in autonomous drone systems using multimodal sensor data.

University Award for Best R&D Project

• Highest marks achieved in undergraduate thesis research on distributed graph ML.

Virtusa Award

• Awarded to the student who earned the highest marks for the semester 5 Software Engineering Project.
• Issued by Department of CSE, University of Moratuwa.

Flu Shot Learning Challenge - 18th/4811 (top 0.4%)

• DrivenData machine learning competition.

SLIIT Codefest Datathon - Champions

• Winning team (XGCoders) of nationwide inter-university data mining competition.

iHack 5.0 - Winners

• 24-hour inter-university software development hackathon at University of Colombo.
• Team ODIGOS designed and developed a mobile/web app solving Sri Lanka's tourism issues. First place out of 59 teams.

Lab Assistant: Programming Fundamentals (CS1033)

University of Moratuwa

Private Tutor (Data Structures, ML, Physics & Maths)

Private Tutoring

• Mentored 60+ students.

Conference Presenter

WSO2 Technology Conference 2025

• Technical deep-dive on blockchain architecture, consensus algorithms, and distributed ledger mechanisms.