skip to content
Open for projects & training
A technology lab from Malaysia

Webuildsoftware, SaaS, AI, systemsyoucantrust.

OrynLab is four builders out of Universiti Kebangsaan Malaysia: SaaS products, IT services and consultation, AI research, and hands-on training — built so the people relying on it can see how it works.

Meet the teamOur servicesWatch the film
— Who we are

We'reatechnologylabthatships.Wefindoutwhattheproblemreallyis,buildthethingthatfixesit,andmakesurethepeoplerelyingonitcanseehowitworks.

Our vision

Technology people can rely on, and understand — from the first prototype to the system that runs your business.

Our mission

Build software, systems and AI that work in the real world, and hold every one of them to one rule: it has to explain itself.

Who we work with

  • Businesses

    Products, apps and data systems that run the day to day.

  • Hospitals & research

    Explainable models, commissioned by specialists.

  • Public sector

    Data for national-language AI, and policy work.

  • Campuses & teams

    Hands-on training that makes AI make sense.

— The film

Two minutes.
Everything we do.

From the stars sailors steered by to the training room — what we build, who we build it for, and the one rule behind all of it.

  • 2:17
  • Sound on
  • Captions available
  • SaaS · IT · Consultation · AI research · Training
— What we do

Five services.
One rule.

Products, systems, advice, research and training — and whatever we build has to explain itself to the people who rely on it. That rule runs through all five.

SaaS products

Software you subscribe to, not rebuild — our own products, or one built and run around your workflow.

  • ResumeJe, our own product: AI-assisted resumes with a rules-based fallback for when the model is unavailable, and an auditable review workflow.
  • RouteIQ ranks field-sales visits by expected value, explains every pick in plain language, and lets ChatGPT and Claude read live route plans over MCP.
  • KDStudio: one Android app for a distributed team's attendance, location checks, tasks and payroll.
amjadh · ezairiel · amerudeen

MCP

live route plans ChatGPT and Claude can read

IT services

Web and mobile apps, data pipelines and cloud work — built to survive real, messy data.

  • Full-stack web (Next.js, FastAPI) and Android (Jetpack Compose, Firebase) — from prototype to production.
  • Document extraction across theses, archives and government records at 95% accuracy.
  • A client dataset of 2,013 restaurants across 7 states, collected past production-grade bot detection — the client came back with a second project.
amjadh · amerudeen · ezairiel

95%

extraction accuracy — theses, archives, records

IT consultation

Know what to build, buy or leave alone — before you spend on it.

  • Technology and AI readiness: where it pays off for you, what data it needs, and what to skip.
  • Model and system audits — leakage checks, held-out testing, and the reasons behind each decision.
  • Governance from people who co-authored a national youth policy report presented to MOFA, the UN and ASEAN.
izzul · ezairiel · amjadh

3

policy audiences — MOFA, the UN, ASEAN

AI research

Commissioned studies and explainable models, on your data.

  • Commissioned by HCTM neurology and rheumatology: an EEG-risk study across 288 patients, and a leakage-aware lupus relapse model.
  • SLEcare turns fragmented lupus-nephritis records into explainable kidney-risk predictions inside a clinician-facing workflow.
  • Training data for WaRiSAN, Malaysia's national academia LLM — 250,000+ synthetic QA pairs.
All four of us

0.929

recall — commissioned geriatric EEG study

Training & workshops

Hands-on sessions that make AI make sense — for students, staff and teams.

  • AI literacy: how AI really works — prediction, tokens, and where it fails.
  • Hands-on with modern AI tools — prompting, Skills, MCP, and connecting AI to the tools you already use.
  • Talks, live demos, games and quizzes, for campuses, companies and communities.
All four of us

Hands-on

every session — demos, games and live quizzes

— Training

We build it.
We teach it, too.

Hands-on sessions for campuses, companies and communities, run by the people who build this every day — so AI stops being a black box for your team.

  • Talks
  • Live demos
  • Games
  • Live quizzes
  • Hands-on
See our sessionsBook a session
  • 01

    AI literacy

    How AI really works — prediction, tokens, where it fails — for students, staff or leadership.

  • 02

    Hands-on AI tools

    Working with modern AI: prompting, Skills, MCP, and connecting AI to the tools you already use.

  • 03

    Build with AI

    For technical teams: shipping AI features that hold up, and that can explain themselves.

  • 1 session run so far · photos on the training page

— How we work

From paper
to product.

Every project runs the same four steps. The last one is the one most teams skip.

  1. Step 01

    Read

    We read the paper before we open the editor — the field, the data, what's already been tried.

  2. Step 02

    Find the gap

    Then we look for the gap nobody else is looking at, and pick the hard version of the problem.

  3. Step 03

    Prototype

    We build the prototype that proves it, on the real problem rather than a toy one.

  4. Step 04

    Explain, then ship

    Nothing is finished until a clinician, a policymaker, or the person it affects can see why it decided what it did.

Bring us

— The team

Four builders.
One lab.

Clinical machine learning, software engineering, AI research and data engineering — the four disciplines our practices need, in one room.

Enter character select
— The work, in numbers

0k+

synthetic QA pairs

for WaRiSAN, Malaysia's academia LLM

0

patients studied

commissioned by HCTM neurology

0%

extraction accuracy

theses, archives, government records

0

policy audiences

the youth roundtable report went to MOFA, the UN and ASEAN