AI Engineer | Tech Enthusiast
My journey into AI began with a love for math and programming, a passion I've since turned into expertise as an AI Engineer building production systems that serve thousands of users.
Today, I architect AI-powered search and chatbot solutions while pursuing graduate research in Computer Vision and Edge AI for real-time anomaly detection. My work is driven by the goal of creating cutting-edge solutions for real-world problems, with a particular focus on applications for Southeast Asia and low-resource languages, like Khmer.
This practical focus is paired with academic rigor—my research on privacy-preserving AI governance was recognized with a Best Paper Award at ASIP 2025. I specialize in turning theory into measurable impact, having achieved improvements like a 132% accuracy boost in semantic search for over 6,000 users. My toolkit is built on a strong foundation in Python, FastAPI, cloud deployment, and the DevOps principles I embraced as a self-taught developer.
I’m relentlessly focused on pushing boundaries, learning, and contributing to intelligent systems that shape the future. The dream, inspired by the visionaries who sparked my interest, is not just to use AI but to build it meaningfully.
Built ML models using Supervised, Unsupervised, and Reinforcement Learning.
Designed and trained deep learning models (CNNs, RNNs) for computer vision and sequence prediction tasks.
Built end-to-end ML pipelines using Python libraries (TensorFlow, PyTorch, Scikit-learn
Deployed scalable ML models on AWS, leveraging cloud computing and storage.
Containerized ML APIs and microservices for portable, reproducible deployments.
Managed version control for collaborative ML projects and CI/CD workflows.
Processed and analyzed structured data for feature engineering and model training.
Automated server tasks and ML pipeline orchestration using Bash scripting.
Built a bilingual (English/Chinese) AI-powered customer service chatbot.
Developed using FastAPI, scikit-Learn, and modern web technologies.
Demonstrated intent classification, entity recognition, and multi-turn conversation capabilities for e-commerce customer support.
Built a machine learning system to detect fraudulent transactions using Logistic Regression, Random Forest, and XGBoost.
Developed a REST API with FastAPI for real-time predictions and containerized the application with Docker.
Achieved 95% precision and 90% recall using SHAP for model interpretability.
Built MobileNetV2 classifier (TensorFlow) with data augmentation (PIL/OpenCV).
Applied transfer learning for efficient binary classification.
Built CNN (TensorFlow/Keras) for cats/dogs classification (85% accuracy).
Optimized with EfficientNetB0 transfer learning and OpenCV preprocessing.
Segmented housing markets using K-means (12 clusters) with dynamic analysis.
Engineered features (population density) and visualized insights with Seaborn.