Proving it can be built here

Inside a regulated Pakistani bank, a team of young engineers I hired in Islamabad has taken thirty AI initiatives from idea to build, eleven of them live, on open models the bank runs itself, with no foreign vendor in the loop. Voice banking, a bank-wide copilot, AI email triage, a financial-crime stack. It is a working answer to the question I hear most often: can we really do this ourselves?

Banking in the languages people speak

Voice banking in Urdu, English, Punjabi and Pashto, across 14 everyday tasks, for people who cannot or would rather not type. More than 20,000 people use it every day, and every one of the bank's 2.3 million customers can. Built in-house, with no vendor between the customer and the bank.

Training the next AI team

I designed and run a paid AI internship that now draws more than a thousand applicants a year. Every candidate sits a real, scored technical assessment, and the strongest spend twelve weeks shipping production work alongside the team.

A decade in the classroom

Before banking I spent almost a decade on the faculty at NUST. I rebuilt the machine learning curriculum around industry projects, supervised MS theses, and mentored students into industry roles and graduate programmes. NUST gave me its Excellence in Teaching Award in 2021, and today I also teach an AI cohort at Edversity.

Research for Pakistan's health

From the Data Analytics Lab I led at NUST: machine learning for sepsis in ICU patients, cancer genomics, and drug discovery, published openly and cited 559 times worldwide. A tutorial I co-wrote on molecular docking is still how many students learn the method.

In the room where it is decided

AI in Pakistan moves when the people at the top see it working, so I put it in front of them. I am Secretary to the President & CEO's “Bank of the Future” forum, I hosted our Board at the National Science and Technology Park, and our team won Fastest to Impact at the Advantage AI Pakistan masterclass.