How I got here
My name is Haoxuan Dong, and I also use the name Jason. I am currently studying in Melbourne, Australia, while using my time outside school for software development, AI experiments, and independent research.
I began using AI tools systematically in September 2025, first as a way to learn unfamiliar concepts and organize information. By January 2026, I was working with agent-based systems that could handle files, call tools, run code, and continue toward a goal over time. That shift made me interested not only in what models can generate, but in why complete technical systems can be reliable.
Moving to Australia made this exploration more accessible. Barriers around services, documentation, open-source projects, and payment became smaller, shortening the distance between “I want to try this” and “I can begin.”
How I work
I usually begin a project with an idea, then ask whether it is worth the investment, what I can realistically take on, and whether it can become a genuinely complete result. AI makes it easier to enter unfamiliar fields and move an idea beyond the edge of what I already know.
I welcome AI in analysis, programming, and checking, but it has also made one thing increasingly important to me: being able to produce a result with AI is not the same as genuinely understanding it. I want to be able to explain a project’s core logic, key judgments, and verification process—and take responsibility for its final result.
What I’m exploring
I have not yet decided whether my future direction will be software, computing, or medicine. I hope to study at the University of Melbourne and remain interested in work at the intersection of technology and medicine.
MidnightDesk
MidnightDesk is where I keep this work and its surrounding context. It is not a product homepage or an extended résumé: the site collects public projects, research, and notes while also keeping some of the thinking behind them.