The Missing Layer: Why Bangladesh's AI Ambition Depends on One Specific Kind of Professional

Bangladesh does not have a shortage of IT workers. Online freelancing alone supports some 650,000 people and brought in an estimated USD 725 million in FY2024-25, and BASIS counts 2,700 member firms behind them.

What it lacks is a layer. AI integration, data engineering, model deployment, cybersecurity: the capabilities commanding rising rates globally are exactly where Bangladeshi supply is thinnest. Volume without that layer means competing on price in the work automation reaches first.

There is a clock. Bangladesh is still scheduled to leave Least Developed Country status on 24 November 2026, and a request to extend the preparatory period awaits a UN General Assembly decision before that date. However it lands, garments remain over 80 percent of exports and roughly three-quarters of merchandise exports currently enter foreign markets duty-free under LDC preferences. Digital services are the designated alternative — but enlarging the volume of subcontracted work would repeat the garment trajectory in a newer industry, and a position built on cheap labour disappears the moment a cheaper supplier appears.

What is needed is a change of position. That takes policy, capital and infrastructure — and something none of those supply. A data center runs only if someone can operate it; a strategy leaves the page only if someone can implement it. The binding constraint is people, the one constraint a country cannot import.

The AI-Hub Project is built around that constraint. Funded by a USD 13 million grant from the Korea International Cooperation Agency (KOICA) under the Record of Discussion signed this week, the center will run advanced curricula across seven tracks through 2029 on renovated floors at Software Technology Park-2 in Karwan Bazar, operated by the Bangladesh Hi-Tech Park Authority under the ICT Division of MOPTIT. Admission requires neither a job nor a degree: a mid-career professional, a fresh graduate and a self-taught freelancer enter through the same door, screened on capability and commitment.

That design supports a national target. The Five-Year Strategic Framework calls for 1.5 to 2 million ICT professionals by FY2031, against roughly 750,000 today. Reaching that number will take more than fresh graduates: it also means converting people already in the workforce into specialists, through structured, credentialed programmes rather than one-off workshops. That is what KOICA has funded the AI-Hub to deliver.

The seven tracks are AI, taught as six-month foundation and advanced courses, alongside big data, cloud computing, smart factory, cybersecurity, blockchain and 3D modelling — core technologies of the Fourth Industrial Revolution, spanning both the infrastructure AI runs on and the industries it lands in. Together they will produce 1,075 direct beneficiaries by 2029 — specialists trained to the level the field demands internationally, in the categories where employers currently cannot fill positions at all.

What a graduate leaves with begins with the capability itself, benchmarked against what the field demands globally rather than what happens to be available in Dhaka. The curriculum is built as compressed, intensive training benchmarked to current international industry practice, and runs that way across all seven tracks. High performers can also sit for national certification — a credential that travels with its holder in a market where skill is easy to claim and hard to verify.

From there the paths fan out. Graduates are connected through a Global Alliance that applies AI-based matching, putting verified capability in front of employers looking for exactly that, at home and abroad. Those who would rather build than be hired can take the startup route with incubation support behind them, others join applied research at the center, and every year a national technology competition puts their work in front of the institutions and firms that might use it.

The KOICA grant also covers training the trainers, with annual workshops and conferences keeping instructors and course content on where the technology is going rather than where it stood when a syllabus was approved. A center that teaches only students stops the day the grant ends; one that leaves behind qualified instructors keeps going.

KOICA has also built in a way to find out whether it worked. An impact evaluation will test, against baseline data and comparison groups, how far participants' capabilities improved and what happened to their employment and earnings after. Attendance figures are the easiest metric to produce and the least informative to read. The design allows for a second phase, which can proceed if those results verify performance sufficient to justify it — so the evaluation is not a report filed at the end but the bridge to whatever comes next.

Bangladesh is the one making the transition — through its infrastructure, its policy, its investment, its firms. What KOICA has put USD 13 million behind is the specialist layer that transition requires: 1,075 people trained to international standard in the fields where Bangladeshi employers cannot fill positions today. Tomorrow's article turns to the other half of the design: what those graduates are meant to build, and for whom.