I’m Trần Tuyến, a pediatric surgeon based in Ho Chi Minh City, Vietnam. I work in neonatal and pediatric surgery — congenital anomalies, neonatal emergencies, long-term outcomes.
I started this blog because I spent years doing research the hard way.
I graduated in 2017, went straight into a surgical residency where I was operating as a primary surgeon from day one, earned my Master’s and specialty qualification along the way, and have been a practicing pediatric surgeon since 2021. I have published in Vietnamese journals, have three papers currently under review at international journals, three more being prepared for submission, and am running two active cohort studies — one clinical, one on AI use in research.
None of that came with a research support system. No funding. No dedicated mentor. No protected research time. Just a full clinical schedule and the kind of questions that don’t let you sleep.
In Vietnam’s medical training system — and in many healthcare systems across the developing world — supervisors are excellent clinicians but rarely trained methodologists. You learn to operate. You are not taught to design a study, navigate peer review, or structure an argument for an international journal. You figure it out alone, or you don’t figure it out at all.
AI entered my workflow not as a shortcut, but as something closer to a thinking partner. I used it wrong at first — everyone does. The hallucinated references, the overconfident summaries, the prose that reads well but means nothing. I made those mistakes, corrected them, and started again. What emerged, slowly, was a set of workflows that actually held up: for reading strategically, for developing arguments, for stress-testing interpretations before reviewers do.
This blog exists to document those workflows — not as rules, but as methods developed through real failures in real academic work.
The reason I think this matters goes beyond productivity. Clinicians in under-resourced settings produce important research. Their observations are often closer to the ground truth of how disease behaves in real populations. But without the methodological infrastructure that well-funded institutions provide, that knowledge stays local, unpublished, invisible. If AI can reduce that gap — not by replacing judgment, but by making the process less opaque — then the patients at the end of that research chain benefit too. That’s the argument I keep coming back to.
What this site is
Three sections, each built around a different layer of academic work:
- Foundations — how the publishing system actually works: what editors optimize for, why good papers fail, how reviewers read.
- Practice — real workflows from question formation to submission, including where things break down.
- AI Tools — specific tools, specific stages, specific limitations. Not reviews for clicks.
Everything here is based on direct experience. The tagline is honest: AI assists thinking — you own the science.
Background
- Medical Doctor; Pediatric Surgeon — Ho Chi Minh City, Vietnam
- Clinical focus: congenital anomalies, neonatal surgery, microsurgery, long-term outcomes
- Research: clinical trials, retrospective cohorts, diagnostics, AI in research
- Epidemiology in Public Health Practice — Johns Hopkins University
- Summary Biostatistics in Public Health — Johns Hopkins University
- Statistics: R (tidyverse, meta), Python (pandas, numpy), jamovi, SPSS
For questions or collaboration: aiforacademic@gmail.com