6,610 career entries from 1,171 alumni across 6 South African private schools, mapped to Karpathy’s 342 BLS occupations. Tile area proportional to career entries. Colour = his exact AI exposure score (0–10).
Data source — Alumni career data sourced from People Data Labs (PDL) Person Search API, March 2026. PDL aggregates professional profile data from LinkedIn and other public sources. 1,171 alumni profiles were identified across 6 South African private schools; each profile’s full career history was extracted, yielding 8,329 individual career entries covering all roles held — not just current positions. For questions about this research: david@cardona.capital.
AI exposure framework — Scores and colour scale follow karpathy.ai/jobs exactly. Andrej Karpathy scored all 342 Bureau of Labor Statistics occupations on a 0–10 scale using an LLM that estimates how much current AI will reshape each occupation, weighting heavily whether the work is fundamentally digital. 0–4 (green) = low; 5–7 (amber) = medium; 8–10 (red) = high exposure. This dashboard applies his scores unchanged.
Mapping methodology — Each raw job title was matched to the closest BLS occupation using a keyword lookup table (~130 profession categories). Longer, more specific keys take precedence (e.g. “business development manager” maps before “manager”). Entries that could not be matched with reasonable confidence — generic titles (intern, retired, student) or highly niche roles — were excluded. 18.3% of entries (1,719 of 8,329) excluded on this basis. All scores and category groupings are Karpathy’s verbatim.