-
9th International Conference on Engineering, Applied Sciences, and Technology (ICEAST)
Vientiane, Lao PDR
Deep CNN classification of banana plant nutrient deficiencies across 9 classes (healthy and 8 deficiency types). Evaluated on two public datasets with augmentation strategies. Achieved ~88% F1-score.
-
8th International Electrical Engineering Congress (iEECON)
Chiang Mai, Thailand.
Classified 7 nutrient conditions in black gram using combined old + young leaf images. ResNet50 features fed into MLP outperformed SVM and logistic regression. Achieved 88.33% accuracy.
-
16th International Joint Conference on Computer Science and Software Engineering (JCSSE)
Pattaya, Thailand.
Benchmarked deep CNN transfer learning for nutrient deficiency detection on 4,088 black gram leaf images across 7 treatment classes. ResNet50 outperformed block-based methods and human-reported baselines. Best model: ResNet50 at 65.44% accuracy.
About Me
Researching efficient vision models.
Scaling them on TPUs.
I'm a PhD researcher at KMITL investigating Knowledge Distillation and Model Compression for Computer Vision, working toward a custom lightweight model that achieves significant size reduction without sacrificing accuracy. My work centers on lightweight, attention-based vision models and how they behave under class imbalance.
As a member of Google's TPU Builders program, I use Cloud TPUs to train and evaluate these models at scale, and I share what I learn through technical tutorials and code for the community. My published IEEE research applies deep learning to agricultural AI.
I also spent 4+ years as an AI/ML Engineer at Six Atomic, building the pattern grading algorithms at the core of the product (cutting manual workload by 90%), AWS-deployed pipelines, and RAG + LLM applications. That experience means I understand both what research claims and what production demands. Outside research I speak at Women in AI Myanmar and mentor at GDG Yangon's ML Paper Reading Club, because the best way to understand something deeply is to teach it.
Beyond the screen
Cat lover Coffee enthusiast Yogini 10k steps daily Nature enthusiast