Infrastructure Bottlenecks and Data Scarcity: Challenges for AI and Machine Learning Engineering Students in Pakistan

Pakistan has recently adopted a National Artificial Intelligence Policy and is expanding AI-related education, yet students who train as AI and machine learning (ML) engineers still work in conditions that differ sharply from those assumed by most published curricula and tutorials. This perspective article examines two interlinked constraints: (i) infrastructure bottlenecks, namely limited access to GPU compute, slow and unevenly distributed connectivity, and power and cost barriers; and (ii) data scarcity, particularly for Urdu and other Pakistani languages and for locally relevant domains such as agriculture, health and flood monitoring. Drawing on policy documents, published research on low-resource language processing, and publicly reported connectivity benchmarks, the article argues that these constraints compound one another and shape what students can learn, build and publish. It then outlines practical mitigation strategies at the level of the student, the institution and national policy. The article is a structured synthesis rather than an empirical study; it does not report a new survey, and its claims about student experience are framed as hypotheses for future empirical work.