06/05/2026
๐ฆ๐ผ๐น๐ผ ๐ฏ๐๐ถ๐น๐ฑ. ๐ฎ๐ฐ ๐ต๐ผ๐๐ฟ๐. ๐ช๐ผ๐ฟ๐น๐ฑ ๐๐ฎ๐ป๐ธ ๐๐ต๐ฎ๐น๐น๐ฒ๐ป๐ด๐ฒ ๐ฌ๐ฑ. ๐ข๐ป๐ฒ ๐ฏ๐ถ๐ป๐ฎ๐ฟ๐, ๐๐๐ผ ๐ฐ๐ผ๐๐ป๐๐ฟ๐ถ๐ฒ๐, ๐ญ,๐ฐ๐ต๐ฌ ๐๐ฆ๐๐ข ๐๐ธ๐ถ๐น๐น๐.
This past weekend I shipped ๐จ๐ก๐ ๐๐ฃ๐ฃ๐๐ โ ๐ฆ๐ธ๐ถ๐น๐น๐ ๐๐ป๐ณ๐ฟ๐ฎ๐๐๐ฟ๐๐ฐ๐๐๐ฟ๐ฒ at the ๐ฑ๐๐ต Hack-Nation's Global AI Hackathon โ the Massachusetts Institute of Technology-backed, 24-hour global AI sprint that pulled in 1,500+ builders from 60+ countries.
I picked the World Bank Group Youth Summit track because the numbers in my own country are hard to look away from: ๐ฏ๐ญ.๐ณ% of Pakistani youth are NEET (not in education, employment, or training). ๐ณ๐ญ.๐ฑ% of non-agricultural employment is informal. Fixed broadband sits at ๐ญ.๐ฐ ๐ฝ๐ฒ๐ฟ ๐ญ๐ฌ๐ฌ. Meanwhile, ESCO โ the EU's open skills taxonomy with 1,490 skills and 411 occupations โ sits in an API that almost no LMIC youth ever touches.
๐จ๐ก๐ ๐๐ฃ๐ฃ๐๐ ๐ถ๐ ๐๐๐ผ ๐๐๐ฟ๐ณ๐ฎ๐ฐ๐ฒ๐ ๐ผ๐ป ๐ผ๐ป๐ฒ ๐ฏ๐ถ๐ป๐ฎ๐ฟ๐:
๐ชช ๐ฌ๐ผ๐๐๐ต ๐๐ถ๐ฑ๐ฒ โ A teenager types their life in free text (Urdu or English). Out comes an ESCO-grounded, portable Skills Passport with QR and JSON-LD. Verifiable. Cross-border. Real.
๐ ๐ฃ๐ผ๐น๐ถ๐ฐ๐๐บ๐ฎ๐ธ๐ฒ๐ฟ ๐ฑ๐ฎ๐๐ต๐ฏ๐ผ๐ฎ๐ฟ๐ฑ โ NEET rates, sector employment, 15-year cohort projections (Wittgenstein SSP2), and skill-supply-vs-demand divergence diagnostics โ wired live to World Bank WDI, ILOSTAT, and the Wittgenstein Centre WCDE-V3.
๐ ๐๐ผ๐๐ป๐๐ฟ๐ ๐ฃ๐ฎ๐ฐ๐ธ ๐ฎ๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐๐ฟ๐ฒ โ Same binary serves ๐ต๐ฐ Pakistan (Urdu) and ๐ฌ๐ญ Ghana (English) with ๐๐ฒ๐ฟ๐ผ ๐ฐ๐ผ๐ฑ๐ฒ ๐ฐ๐ต๐ฎ๐ป๐ด๐ฒ๐. New LMIC = config swap, not a rewrite. Deployed on AWS Amplify.
๐ ๐๐ถ๐๐ฒ: https://lnkd.in/dNgiPmU3
๐ ๐๐ผ๐ฑ๐ฒ: https://lnkd.in/dH3HFrPH
๐ฆ๐ผ๐น๐ผ ๐ณ๐๐น๐น-๐๐๐ฎ๐ฐ๐ธ โ design, ESCO API integration, free-textโskills mapping, multilingual UI, policy dashboard, deployment. Ship beats perfect. Always.
Massive respect to Kai Nestor Wiederhold and Linn Bieske for orchestrating a global build sprint that takes problems like this seriously, and to Hack-Nation's Global AI Hackathon, MIT Sloan AI Club, MIT Club of Northern California, and MIT Club of Germany for backing builders working on more than another ChatGPT wrapper.
๐ง๐ต๐ฒ ๐๐ฎ๐ธ๐ฒ๐ฎ๐๐ฎ๐ ๐ฎ๐ณ๐๐ฒ๐ฟ ๐ฎ๐ฐ ๐ต๐ผ๐๐ฟ๐: AI's greatest leverage isn't in the model โ it's in the ๐ถ๐ป๐ณ๐ฟ๐ฎ๐๐๐ฟ๐๐ฐ๐๐๐ฟ๐ฒ ๐๐ฒ ๐๐ฟ๐ฎ๐ฝ ๐ฎ๐ฟ๐ผ๐๐ป๐ฑ ๐ถ๐ ๐ณ๐ผ๐ฟ ๐๐ต๐ฒ ๐ฝ๐ฒ๐ผ๐ฝ๐น๐ฒ ๐๐ต๐ผ ๐ป๐ฒ๐ฒ๐ฑ ๐ถ๐ ๐บ๐ผ๐๐.
๐ฆ๐ธ๐ถ๐น๐น๐ ๐ถ๐ป๐ณ๐ฟ๐ฎ๐๐๐ฟ๐๐ฐ๐๐๐ฟ๐ฒ ๐ถ๐๐ป'๐ ๐๐ฒ๐
๐. ๐๐ ๐ท๐๐๐ ๐ฑ๐ฒ๐ฐ๐ถ๐ฑ๐ฒ๐ ๐๐ต๐ผ๐๐ฒ ๐๐ฎ๐น๐ฒ๐ป๐ ๐ฐ๐ผ๐๐ป๐๐.
hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag hashtag