Source-linked AI summary
VietProfs: A Public Directory of the Vietnamese Academic and Research Diaspora
ThanhVu Nguyen
TL;DR
Vietnamese-diaspora scholars are spread across institutions without a centralized directory, and keeping records current against changing careers and websites is difficult. VietProfs combines an open searchable directory with agent-assisted discovery and verification, bounded by deterministic checks. Its September 2026 snapshot contains 1,152 verified scholars across 492 institutions in 23 countries, while its counts are not a demographic census.
Problem
Vietnamese-diaspora scholars lacked a centralized searchable directory, while changing appointments and institutional websites make manual record maintenance difficult.
Source: §1 · p. 1, §1.2 · p. 3Method
VietProfs combines an open searchable directory with automated expansion and revalidation, using AI for proposals and deterministic checks to validate records.
Source: §1 · p. 2, §1.1 · p. 2, §1.2 · p. 3Results
As of September 2026, VietProfs includes 1,152 verified scholars across 492 institutions in 23 countries.
Source: §8.1 · p. 10Takeaways & Limitations
The directory provides a practical resource for finding mentors and colleagues and documenting diaspora contributions across generations.
Source: §20 · p. 16VietProfs is a curated directory, not a demographic census; coverage depends on institutional web transparency, English-language profiles, and search effort.
Source: §9 · p. 11
Abstract
from arXiv · showhide
The Vietnamese academic diaspora spans hundreds of universities and research institutes worldwide, yet there has never been a shared, searchable directory for this community. Prospective graduate students looking for advisors who share their background, researchers seeking collaborators, and organizers looking for speakers have had to rely on manual, one-off web searches. To address this need, we built VietProfs (https://vietprofs.roars.dev), an open, searchable directory of Vietnamese and Vietnamese-diaspora faculty and permanent research scientists outside Vietnam. Each profile pairs a portrait and authoritative diacritic name with verified appointment, degree, and honors data, browsable through a fast client-side web interface. This paper covers two topics: the design of the directory itself and the engineering required to keep it current. Part I describes the directory: its goals, eligibility and appointment track rules, alignment with standard U.S. federal classifications (NCES, NSF, and NIH), profile card design, and current roster insights. Part II reports on automated directory maintenance. Keeping over a thousand records accurate against a constantly shifting web is notoriously difficult. We describe an LLM-assisted maintenance system combining automated search and periodic revalidation with deterministic validation gates, operating under the principle: AI proposes, evidence decides. As of September 2026, the directory includes 1,152 verified records across 492 institutions in 23 countries, covering doctoral cohorts from 1939 to 2026. Aggregate statistics are computed dynamically at runtime, and all changes are audited through public Git history. VietProfs provides a practical community resource while offering a concrete case study in building reliable, agent-maintained public datasets.
1 Introduction
VietProfs addresses the lack of a centralized, searchable directory for Vietnamese and Vietnamese-diaspora scholars abroad. It pairs a curated directory with automated workflows intended to keep records accurate as careers and institutional websites change.
1 Introduction
- The diaspora lacked a centralized, searchable record, leaving students, researchers, and organizers to find scholars through ad hoc searches.Informal rosters in other diaspora communities have supported mentorship, collaboration, and knowledge sharing.Source: §1 · pp. 1–2
1.1 What VietProfs Is: Motivation, Goals, and Design
- VietProfs is an open directory designed for mentor searches, cross-institution connections, event planning, and documenting diaspora achievements.Profiles include verified appointments and degrees, honors, research tags, and authoritative Vietnamese-diacritic names; the interface offers search, filters, and shareable URLs.Source: §1 · p. 2, §1.1 · p. 2
- Eligibility requires a verified appointment outside Vietnam in an accepted track, while temporary, visiting, postdoctoral, and adjunct roles are excluded.The Research track includes permanent faculty-equivalent researchers at public bodies and research institutes.Source: §1.1 · p. 2
1.2 The Engineering Challenge: Keeping VietProfs Accurate
- Keeping records accurate is the central engineering challenge because careers change and institutional pages can redesign or break.The system uses continuous expansion to discover eligible scholars and revalidation to check existing records; community-submitted leads are useful for discovery, while agents verify claims against official pages and humans review high-impact edits.Source: §1.2 · p. 3
Part I: VietProfs: Purpose, People, and Content
VietProfs provides a shared directory for finding Vietnamese-heritage scholars and exploring the diaspora’s academic reach. Its scope is verified scholars identified by the project, not a complete census or ranking.
2 Goals and Intended Uses
- VietProfs offers a centralized answer to where Vietnamese-heritage scholars can be found, addressing a gap previously met by one-name-at-a-time searches.The shared record spans generations, from early graduates to current assistant professors.Source: §2 · pp. 3–4
- The directory supports mentor searches, cross-institution connections, event organization, documentation of contributions, and exploration of geographic and disciplinary patterns.Users can search by field and location rather than guessing names from departmental rosters.Source: §2 · pp. 3–4
- The roster is neither a scholar or institution ranking nor a complete census; its coverage reflects public web visibility and search coverage.It represents scholars identified and rigorously verified, not the absolute population total.Source: §2 · p. 4
3 Who Is Included: Eligibility
VietProfs includes scholars with verified eligible appointments outside Vietnam and applies track rules intended to keep the roster stable. It uses names only as discovery leads, not as proof of heritage.
3.1 The core requirement
- Inclusion requires reliable evidence of an eligible academic appointment outside Vietnam in one of seven accepted tracks.The geographic boundary addresses a diaspora scattered across institutions worldwide, rather than making a value judgment about Vietnamese universities.Source: §3.1 · pp. 4–5
3.2 Who is not included, and why
- Temporary and term-limited roles, graduate teaching assistants, and industry-only positions are excluded to limit turnover and stale entries.Instructor and Research Assistant Professor titles are assessed case by case, and Professor of Practice roles are included under the Teaching track when stability is confirmed.Source: §3.2 · p. 5
3.3 Identity is not inferred from a name
- Common Vietnamese names are discovery leads, but name alone is neither proof of heritage nor a basis for requiring documentary proof.Surname-based signals can produce false positives and miss scholars; people who believe they are misidentified can request removal or correction through the public form.Source: §3.3 · p. 5
4 Fields of Study Covered
VietProfs organizes scholars into fields adapted from U.S. federal classifications and designs profile attributes to support recognition, discovery, and verification.
4 Fields of Study Covered
- VietProfs’s 17 broad fields adapt U.S. federal classification systems, including NCES’s CIP and NSF’s Survey of Earned Doctorates.The correspondence is intended to make categories familiar and connect the directory to national doctoral-education statistics.Source: §4 · pp. 5–6
- Table 2 maps VietProfs fields to CIP families, combining related areas such as Humanities and grouping Economics with Business for browsing.The Economics–Business grouping reflects that faculty in those areas are frequently hired and housed together.Source: §4 · p. 6, Table 2 · p. 6
- Health Sciences comprises 228 records, roughly 20% of the roster, including 104 Clinical appointments, and is divided into seven NIH-aligned subfields.The subfields correspond to NIH institute structure.Source: §4 · p. 6, Table 3 · p. 6
5 What Information Is Shown, and Why
- Profile attributes support identity recognition, filtering, mentor discovery, scholarly context, and verification through portraits, diacritic names, appointment details, education, honors, research tags, and links.A visible verification timestamp gives visitors a measure of data freshness.Source: §5 · p. 7
6 Data Model and Curated Attributes
VietProfs stores a structured public roster and restricts honors to validated distinctions. Its client-side interface supports searching, browsing, and community-submitted corrections.
6 Data Model and Curated Attributes
- Each roster record stores identity, appointment, location, research, education, honors, links, portrait provenance, and verification date; review timestamps live in a separate ledger.The public source of truth is public/data.json, while maintenance/verification.json tracks completed full reviews by canonical name.Source: §6 · p. 7
6.1 Curated Honors and Awards
- Stored honors must use an accepted category and include a name, granting organization, HTTPS provenance URL, and valid calendar year.The honor categories are restricted to substantial distinctions rather than routine CV entries.Source: §6.1 · p. 8, Table 5 · p. 8
7 User Interface and Interaction Model
- The client-side interface combines searchable profile cards with location, field, and appointment-track filters.Cards show portraits, English and diacritic names, appointment details, education, honors, research tags, and a verification date.Source: §7 · pp. 8–9, Figure 1 · p. 8
7.1 Scoped Search and Diacritic Normalization
- Search matches names and profile attributes through a diacritic-normalized index, while scope controls and URL-synced filters make queries precise and shareable.For example, typing “Nguyen” matches a diacritic form such as “Nguy˜en.”Source: §7.1 · p. 9, Figure 2 · p. 9
7.2 Leaderboards, Geographic Map, and Dynamic Facts
- Clickable leaderboards support institutional drill-down, and the U.S. choropleth map shades states by faculty count and filters the roster when a state is selected.The map appears after selecting a U.S. location or choosing “Show me something interesting.”Source: §7.2 · p. 9, Figure 3 · p. 10
7.3 Human Contribution and Corrections Channel
- Visitors can propose profiles or corrections with a name and verifiable link, but submissions enter review and cannot directly change the canonical database.The form can pre-fill an existing record to support a correction rather than a duplicate.Source: §7.3 · pp. 9–10
8 What the Directory Reveals
The September 2026 snapshot makes the diaspora visible across institutions, countries, and fields. Runtime aggregation and an 87-year doctoral-date span add current and historical views of the roster.
8.1 Runtime Aggregation and Current Distribution
- Aggregate metrics are recomputed from the live dataset at render time rather than stored as static statistics.When primary records change, dependent statistics update automatically.Source: §8.1 · p. 10, §8.2 · p. 11
- 1,152 verified scholars span 492 institutions in 23 countries; the United States has 729 records and the international subset has 423.California (148) and Texas (98) lead U.S. counts; Australia (95) leads the international subset, followed by Canada (58), the United Kingdom (57), France (46), Japan (40), and Singapore (27). The largest broad fields are Health Sciences (228), Business & Economics (220), Computer & Information Sciences (163), and Engineering (152).Source: §8.1 · p. 10, Table 6 · p. 11
8.2 Historical Breadth and Multi-Generational Legacy
- Recorded doctoral graduation dates span 87 years, from 1939 to 2026, documenting a multi-generational presence abroad.The passage cites Buu-Hoi Nguyen-Phuc’s 1939 Paris doctorate and Xuong Nguyen-Huu’s 1962 UC Berkeley PhD as early examples.Source: §8.2 · p. 11
9 Limitations and Ethical Considerations
VietProfs is a curated public directory, not a demographic census. Its counts and coverage are bounded by public-web discoverability, and it limits its data to public professional information.
9 Limitations and Ethical Considerations
- Directory counts should not be treated as demographic totals, migration paths, or measures of institutional prestige.The paper characterizes VietProfs as a curated directory rather than a complete census.Source: §9 · p. 11
- Coverage depends on institutional web transparency, English-language profile availability, and search effort.These factors constrain who is discoverable in the directory.Source: §9 · p. 11
- The directory uses publicly available professional data and avoids invasive ethnicity tests, private contact information, and unverified claims.Source: §9 · p. 11
10 Why Maintaining the Directory Is Hard
Maintaining the directory is difficult because the population is globally distributed and public sources vary in reliability, terminology, and persistence. The system routes discovery and revalidation through evidence review and deterministic gates before canonical changes are committed.
10 Why Maintaining the Directory Is Hard
- Five challenges are open-world discovery, heterogeneous sources, non-standard titles, name disambiguation, and link rot that can mimic departures.The population has no central registry, and a 404 must be distinguished from genuine departure or retirement.Source: §10 · p. 12
11 System Architecture
- Candidate discovery and periodic revalidation feed evidence to LLM agents, but proposed patches must pass deterministic validation before changing the canonical repository.The gates cover schema, appointment track, URL, and target scope.Source: §11 · p. 12
11.1 Controller Execution and Proposal Validation
- The controller selects stale or unverified records and accepts proposals only when schema, target-scoping, and provenance invariants hold.Its stale-record threshold defaults to 365 days; honors require valid HTTPS sources, accepted categories, and non-future years.Source: §11.1 · pp. 12–13
11.2 Independent Review and Resumable Checkpoints
- A second model reviews patches against web evidence; unresolved discrepancies are deferred after at most two revisions, while saved working state supports resumption.The review is enabled with --codex-review, and state persists across restarts or rate limits.Source: §11.2 · p. 13
12 Discovery as Partitioned Open-World Search
VietProfs bounds open-world discovery with targeted searches, then treats leads as hypotheses until public evidence verifies them. Periodic revalidation checks existing records through a separate maintenance loop.
12 Discovery as Partitioned Open-World Search
- Partitioned institution, department, and domain searches make open-world candidate discovery bounded and repeatable.Queries combine institutional and directory terms with Vietnamese name tokens and negative filters, then check leads against the roster before full verification.Source: §12 · pp. 13–14, Figure 6 · p. 14
13 Evidence and Verification
- Candidate discovery is separate from verification: official university faculty pages are primary evidence, while personal sites, Google Scholar, and award announcements corroborate details.Model outputs remain structured hypotheses until validated against web provenance under “AI proposes, evidence decides”.Source: §13 · p. 14
14 Continuous Revalidation and Self-Maintenance
- Periodic revalidation checks profile liveness, URL roles, scholar identity, appointment and rank, education and honors, then records a timestamped ledger update.Figure 6 depicts expansion and revalidation loops; both newly discovered candidates and stale entries undergo evidence verification before canonical commitment.Source: §14 · p. 14, Figure 6 · p. 14
15 Testing and Validation
Maintenance is guarded by automated tests, audited through representative Git-history events, and informed by shortcuts that failed in practice. The resulting lessons emphasize bounded, evidence-backed proposals and safeguards around canonical data.
15 Testing and Validation
- Every commit and pull request runs data-invariant, unit, route-integrity, and end-to-end browser tests.The checks cover schema and URL rules, field mapping and search behavior, stable profile routes, filters, and submission flows.Source: §15 · p. 14
16 Operational Experience and Evaluation
- Git-audited events illustrate how proposals, deterministic gates, and human oversight handled duplicates, track corrections, affiliation moves, and negative searches.Table 9 documents representative maintenance events from the public Git history.Source: §16 · p. 15, Table 9 · p. 15
17 What Does Not Work
- Unconstrained model queries hallucinated appointments, surname-only filtering misclassified scholars, 404-based removals deleted active faculty, and direct model writes corrupted records.The surname shortcut both admitted false positives and missed Vietnamese scholars with non-Vietnamese surnames.Source: §17 · p. 15
18 Practical Engineering Lessons
- The first five lessons require deterministic invariants, staged leads, target-scoped edits, runtime aggregates, and corroboration before removals.A broken link alone is not sufficient evidence to remove an entry.Source: §18 · p. 15
- The remaining lessons call for a two-tier verification ledger, curated honors, partitioned search, bounded cross-checking, and end-to-end production tests.These practices distinguish audit dates from modification dates and protect production with automated checks.Source: §18 · p. 16
19 Related Work
VietProfs connects appointment verification with existing work on entity resolution and scholarly information systems while using federal educational taxonomies. The paper concludes that an evidence-backed directory can be maintained over time and outlines practical extensions.
19 Related Work
- VietProfs applies entity-resolution and LLM matching ideas [4] [14] [8] to appointments, complements scholarly indexes such as OpenAlex [15], DBLP, and ORCID, and uses federal taxonomies.Its focus is active faculty appointments, tracks, and diaspora visibility rather than publication and citation networks.Source: §19 · p. 16
20 Conclusion and Future Work
- VietProfs pairs agent research with deterministic validation to maintain over a thousand records without an army of human curators.The conclusion describes the directory as an authoritative record of diaspora reach and a practical resource for students and researchers.Source: §20 · p. 16
- Planned extensions include academic genealogy, proximity-based mentorship matching, collaboration bridges with Vietnam, and scholar self-service verification.Self-service updates would retain deterministic schema invariants.Source: §20 · p. 16
- The design philosophy remains “AI proposes, evidence decides,” with automated tools anchored to verifiable public evidence.The paper frames this as the basis for keeping community resources reliable, open, and lasting.Source: §20 · p. 16