Arya Raj Khadka

I’m a final-year Mathematics & Computer Science student specializing in Data Science, Machine Learning, and Analytics. I’m driven by applying analytics to real-world industrial and operational environments, with a strong foundation in statistical modelling, optimization, and data-driven decision making.

In my current role as a Research Assistant - Computer Vision for Warehouse at the University of Adelaide, I engineered a production-grade Android application using CameraX, ML Kit, and OpenCV for real-time Data Matrix and AprilTag detection. I built a multi-tag vision pipeline (4-point corner localization and ID extraction) to enable automated pallet and box tracking, working closely with academic researchers and industry partners to translate operational needs into deployable solutions.

Across research and competitions, I’ve built and validated ML systems end-to-end—from a PyTorch pipeline inferring epidemiological parameters from phylogenetic trees to winning the AIML Grand Discovery Summer Challenge with a multi-modal architecture that achieved final MAE 0.5766. I also develop agentic and deep learning projects, including an NDIS Progress Note Generator Agent using LangGraph (with audio transcription, structured output, human-in-the-loop review, FastAPI endpoints, and a Streamlit interface) and a 3D deep learning CT pipeline for Hemifacial Microsomia severity classification.