EXPERIENCE
INFOCENTRIC
Focused on backend performance engineering and distributed systems architecture, driving a 72% reduction in end-to-end service latency and designing event-driven microservices pipelines for scalable, fault-tolerant data processing.
- Used JFR and .NET Trace to profile backend services under load, identifying CPU bottlenecks and concurrency hotspots across thread execution paths and shared-state contention points.
- Improved parallel execution and reduced contention through multithreading and JVM garbage collection tuning, reducing end-to-end latency from 2 minutes to 33 seconds (72% improvement) while improving system stability under high concurrency.
- Architected an event-driven microservices pipeline in .NET, orchestrating Docker containerised connector, parser, chunking, and analysis services via Redis-backed asynchronous job queues, enabling scalable and fault-tolerant data processing.
- Leveraged Snowflake to build scalable data pipelines and query layered datasets, and developed interactive R Shiny dashboards to transform analytical outputs into stakeholder-facing visualisations for reporting and decision support.
MONASH UNIVERSITY
Working alongside clinical researchers with no software background, I took vague, high-level goals and turned them into a fully designed, production-grade application. That meant asking the right questions, understanding what they actually needed versus what they thought they needed, and making every design decision with the end user in mind.
- Translated non-technical clinical requirements into a full system design for Respio, a cross-platform Flutter application for continuous respiratory monitoring, making architectural and UX decisions that aligned with how researchers and patients would actually use the product.
- Architected and developed the real-time data ingestion layer, streaming physiological sensor data over Bluetooth Low Energy (BLE) at 500 Hz with minimal latency and no data loss across extended monitoring sessions.
- Engineered a scalable cloud-backed data architecture using Firebase Authentication and Cloud Firestore to securely manage user identities and persist structured respiratory recordings.
- Developed a document generation pipeline that converts recorded physiological signals into structured PDF health reports with statistical summaries and waveform visualisations, designed to be readable by clinicians without any technical background.