🔗 Two centres, one team, starting today.

Effective August 14, 2026, S Technologies' two research centres are one 🎉

S Web Research Center (SWRC) 🔬, opened on June 16, 2016, and S AI Research Center (SAIRC) 🧠, opened on August 14, 2019, are merged into a single independent centre: S Technologies Research Center (STRC) ✨

Same people, same projects, same mandate — in one team instead of two. Nothing was closed and nobody left 💚

🤔 Why now

The two centres were founded three years apart to answer two different kinds of question. SWRC worked on web technology — frameworks, template engines, developer tooling, and the publishing platforms built on top of them. SAIRC worked on machine learning and deep learning, trained for narrow, real problems rather than for benchmarks.

For most of a decade that was a clean line. It stopped being one.

The clearest example is our own work. Shunno, the general-purpose programming language in Bangla, needs a compiler — that is squarely web-centre work in the old division — and it also needs a language model that understands Bangla well enough to be useful in an editor, which is squarely AI work. Fishtock, the fish-stock forecasting model, is a machine learning project that nobody can actually use until it is a web application sitting in front of a fisheries officer.

Both projects had been running across two centres, with the coordination cost that implies. When the interesting questions consistently need both halves, keeping two centres is the harder way to do the work, not the more rigorous one.

🛡️ What is not changing

This is the part that matters most to anyone who depends on something we published.

  • Every released project stays released. Same repositories, same licences, same documentation, same support.
  • The centre is still independent of the service side. STRC does not report to the part of the company that bills clients. That independence is the entire reason a research centre is a separate thing, and it survives the merger intact.
  • Everything is still published with its source. A result nobody can read or check is not a result.
  • Bangla stays first-class. Programming in Bangla, and computing that reads and writes it properly, is a standing line of work — not a localisation task bolted onto something already finished.
  • Narrow problems still beat leaderboards. We would rather be useful on one real problem than competitive on a benchmark.

🔬 What the centre works on

STRC carries six areas forward, and each one has something released behind it rather than a plan:

  1. Languages and compilers — programming in Bangla rather than English, and what it takes to make a general-purpose language usable by people who never learned to code in a second language. Shunno.
  2. Machine learning and deep learning — models for conversation, classification and prediction, small enough to run where our clients already are. Britto AI, Fishtock.
  3. Web frameworks and engines — how much framework a small team actually needs, worked out by building lightweight MVC, template and database layers and then using them on real projects. S PHP Engine, S Template Engine, S Database Explorer.
  4. Developer tooling — the time between an idea and a running project. Shemul CLI.
  5. Desktop and platform software — what a browser is when it is stripped back to the parts people use. Jaal Browser.
  6. Privacy and the open web — consent and data handling as something a small site owner can install, rather than a policy document they have to interpret. CookieCons, ProjectPress.

The first two came from SAIRC, the next four from SWRC, and the merged centre is the first time they sit in one plan.

🧪 How work leaves the centre

The process is unchanged, because both centres already ran it. Five steps, and nothing is published until it has survived all five:

  • Question — the problem in one sentence, with who has it and what they do today instead. If the sentence will not hold up, the work stops here.
  • Survey — read what already exists. Most questions are answered somewhere, and finding that out early is the cheapest result a study can produce.
  • Prototype — the smallest thing that can be wrong in public.
  • Evaluate — used on real work, by us first. What it costs to run, to learn and to maintain is measured before anything is promised to a client.
  • Release — documented, published with its source, and then kept.

🗂️ The old sites

swrc.stechbd.net and sairc.stechbd.net are still up, and each now opens with a notice explaining the merger and pointing at where the work went. We are keeping them rather than redirecting, because a centre that ran for ten years has a public record and that record should stay readable at the address people cited.

The new home is strc.stechbd.net, in English and Bangla, with the research areas, the released projects and the lineage of how the centre came to exist.

💬 Talk to us

Are you a researcher, a student, or an organisation with a problem that looks like one of the six areas above? 👋 The centre is reachable through its own site, and we would genuinely like to hear from you. If you are after commercial work rather than research — an AI model built for your business, or an application around one — that is the service side, and the contact form is the right door.

Ten years of web research and seven of AI research go into this 🚀 The name is new; the work is not.