Project Portfolio Building for Vietnamese Adolescents 14–19: A Five-Category Typology, Operational Scoping Rule, and Audience-Aware Documentation Practice for University Admission

A Vietnamese adolescent in the 14–19 window who has accepted the I4.1 finding that a one-to-three-substantiveproject portfolio is among the six primary pre-university investments still faces an unanswered operational question: what counts as a substantive project, how is one scoped without overshooting, and how is one documented so that the work is legible to evaluators whose criteria differ across audiences. The popular Vietnamese discourse on portfolio building treats project as a single category and more is better as the operational rule, which produces two dominant failure modes: tutorialfollowing portfolios whose volume is high and whose signal is near-zero, and hackathon-only portfolios that demonstrate a sprint capability but no sustained-investment capability. The 2024–2026 generative-AI deployment wave has added a third failure mode — the AI-augmented project presented as original work without disclosure of how AI was used and how the AI output was verified. This paper synthesises the projectbased-learning, signalling, experiential-learning, and portfolioresearch literatures with the 2024–2026 AI-era career-launch evidence to produce an operationally usable specification. The synthesis is anchored in Spence 1973 signalling theory; the Caplan 2018 versus Acemoglu-Autor 2011 / Goldin-Katz 2008 / Chetty et al. 2014/2017 signalling-vs-human-capital debate; project-based-learning (Krajcik-Blumenfeld 2006; Thomas 2000; Kokotsaki et al. 2016); experiential-learning and reflectivepractice (Kolb 1984; Dewey 1938; Schön 1983/1987); autonomymastery-purpose (Pink 2009; Ryan-Deci 2017); career-capital (Newport 2012/2016/2024); supercharging-techniques (Lin Z. 2025); GitHub social-coding research (Marlow et al. 2013; Tsay et al. 2014; Dabbish et al. 2012); MOOC and hackathon-outcomes literature (Reich-Ruipérez-Valiente 2019; Trainer et al. 2016; PeThan et al. 2020); US college-admissions evidence (Selingo 2020); and Vietnamese-context coverage of KHKT, NCKH, VinIF, and FPT Edu hackathon archives. The paper distinguishes five project categories — flagship (50–200 hours over 6–12 months), skill-demonstration (10–30 hours over 1–3 weeks), collaboration (team-based), replication-or-extension (extends published work), and original-investigation (research-oriented) — and argues that the flagship is the load-bearing portfolio element, that multipleswithout-flagship signal tried-many-things not did-something-well, and that the operational scoping rule is minimum viable substantive rather than minimum viable product. The substantive