Practitioner, Not Theorist

I started building digital systems years before "AI" became a marketing word. The work was the same as it is today — replace humans clicking buttons with reliable software that runs while everyone sleeps.

What changed is the toolkit. Make.com, n8n, Notion, Anthropic, and OpenAI made it possible to ship in days what used to take a six-month engineering build. The constraint stopped being "can we build this" and became "do we know what to build."

Most consultants in Bangladesh sell the technology. I sell the outcome — a system that runs, hours saved, money recovered. If a client cannot point to a process that's now invisible to them, I haven't done my job.

BangladeshAI grew out of the same instinct. The country was ranked, surveyed, and benchmarked by international agencies — but no one was translating those numbers into operator-level decisions for local businesses. So I started publishing the data, with the gaps mapped and the levers identified.

Most businesses in Bangladesh run on human bandwidth. People are acting as APIs — copying data from one system, formatting it, and pasting it into another. This is fragile, unscalable, and ultimately limits the growth of the private sector.

I believe in deterministic systems for reliable tasks (invoicing, reporting, data movement) and probabilistic models (LLMs) for cognitive tasks (summarization, sentiment analysis, drafting).

Bangladesh doesn't need more conference panels about AI. It needs operators who can wire up Make.com, write custom Python scripts, securely connect local APIs like bKash, and hand over a functioning system that saves 15 hours a week.