MD Rifat Hasan
Software engineering student in Dhaka. I build custom ecommerce and the automation that keeps every order honest.
I write the storefront and the machinery behind it. Product pages, checkout, courier handoff, event tracking, and the admin screens somebody opens at 2am to fix an order.
Most of my work is for Bangladeshi ecommerce, where cash on delivery turns every fake order into a real loss. So I build the checks in rather than bolt them on later.
I study software engineering at Daffodil International University. Outside client work I run a YouTube channel where I break down how AI tools actually work, without the hype.
Custom platform work, platform builds on WordPress and Tilda, and one store I ran through its growth years.
A custom storefront and back office for an AirPods accessories brand in Bangladesh. Catalog, cash on delivery checkout, courier handoff and marketing all run on one Postgres schema I designed.
A quiet, photography led site for a luxury property firm trading since 1985. Multilingual listings, enquiry routing, and a layout that lets the images carry the page.
A mother and baby retailer serving all of Bangladesh. I built and ran the WordPress store through its growth stage. The team later moved the catalog to Shopify.
A cross border commerce agency in Malaysia. I worked on the design revamp: hierarchy, section rhythm, and a homepage that explains the service before it asks for anything.
Three of these run in production on Podsy today. They exist because cash on delivery punishes a store for every fake order, every mistyped address and every purchase event that never really happened.
Cash on delivery means anyone can order with a fake number. Every refused parcel is a return fee the store eats.
A courier history lookup runs the moment an order lands. It pulls that phone number's delivery and return record across couriers, scores it, and flags the risky ones for a confirmation call before anything ships.
The team stops guessing. Risky orders get a phone call, clean orders go straight to packing.
Retyping an address into the courier dashboard for every order is slow, and one typo sends a parcel to the wrong thana.
The full Pathao city, zone and area tree is synced into Postgres, so the admin resolves an address to real courier IDs while it is being typed. One button creates the consignment and writes the tracking code back onto the order.
Dispatch becomes a single click, and courier status flows back into the order timeline on its own.
Browser pixels miss events and fire on orders that later get refused. Meta then optimises for buyers who never actually pay.
Purchase events go server side through the Conversions API, keyed by event ID so they deduplicate against the pixel. The event fires on a confirmed order, not a page view, and refused parcels never get counted.
Ad spend is trained on real deliveries instead of fake orders, which is the difference that decides a cash on delivery business.
A storefront and an admin shaped around how your team actually works, not around a theme's assumptions. Next.js on Postgres, and the code is yours.
Courier handoff, fraud checks, stock alerts, receipts. If somebody retypes it every day, it can run itself.
Server side events, deduplicated against the pixel, fired on confirmed orders so your ad platform learns from real revenue.
Sometimes the right answer is a platform you already pay for. I build and fix on all three without pretending you need a rewrite.
Core Web Vitals, image pipelines, structured data. The unglamorous work that decides whether anyone finds the site at all.
Plain explanations of what these tools do and where they break. No hype, no thumbnails shouting at you. If you have ever wondered whether a model actually understands your prompt, that is the channel.
Watch on YouTubeNew uploads land here automatically. In the meantime the channel is at @rifat66z.
Thirty minutes. Tell me what you are building and I will tell you straight whether I am the right person for it.
For anything that does not need a call. I read every message and reply within a day or two.