I'm a Machine Learning Engineer and Data Scientist with experience developing applied AI and machine learning solutions across healthcare, medical imaging, computer vision, sensor analytics, seismic data, video analysis, and software engineering.
I completed my PhD at the University of Auckland, where my research focused on robust, uncertainty-aware, and scalable deep learning for bi-atrial segmentation from LGE-MRI in atrial fibrillation. My work brings together medical image analysis, representation learning, uncertainty estimation, model calibration, and clinically meaningful evaluation.
Alongside my research, I have worked on end-to-end machine learning systems involving seismic waveform analysis, anomaly detection, video analytics, physiological signals, medical images, audio, and structured data.
- Machine Learning and Deep Learning
- Computer Vision and Medical Image Analysis
- Self-Supervised and Representation Learning
- Uncertainty Estimation and Model Calibration
- Time-Series and Sensor Analytics
- Generative AI, LLMs and Retrieval-Augmented Generation
- Reliable and Scalable Machine Learning Systems
Languages: Python, SQL, MATLAB, Go, JavaScript, Bash, R
Machine Learning: PyTorch, TensorFlow, Keras, scikit-learn
AI & GenAI: LLMs, RAG, LangChain, CrewAI, vector databases, AI agents
Data & MLOps: Docker, Airflow, MLflow, PySpark, Kafka, Git, Linux
Domains: Medical Imaging, Computer Vision, Time Series, Sensor Analytics, Video Analysis, Signal Processing
Developing machine learning pipelines for automated P- and S-wave phase identification across thousands of seismic events and station recordings, including model evaluation, uncertainty analysis, and earthquake-location workflows.
Developed deep-learning methods for segmentation of the left and right atrial cavities and walls, with a focus on robustness, uncertainty, generalisation, and clinically meaningful evaluation.
Research on self-supervised representation learning for video understanding, including work spanning multiple families of video learning tasks.
Developed deep-learning-based video analytics for CCTV systems operating across thousands of cameras, including false-alarm reduction and deployment-oriented model debugging.
My research spans:
- Medical image segmentation
- Self-supervised learning
- Uncertainty-aware deep learning
- Model calibration
- Video understanding
- Semi-supervised learning
- Representation learning
You can find my publications through my Google Scholar.
I'm particularly interested in building reliable AI systems that translate complex real-world data into practical and measurable outcomes.