Ahsan Habib

EducatorEngineerResearcher

Ahsan Habib.

আহসান হাবীব

Three disciplines, one question: how does an idea take hold?

I build learning and assessment software, teach programming at university level and research when machine-learning results deserve our trust.

Explore my work Download CV Say hello
Portrait of Ahsan Habib
Kuala Lumpur GMT+8Rooted in Bangladesh
  1. Teaching since2013
  2. Researching since2016
  3. Building since2018
  4. PearlHelix founded2026
Scroll

01The connecting thread

Three strands. One thread.

I trained in electrical engineering and computer science at North South University in Dhaka, then came to Universiti Malaya to research how people learn to think computationally. I never stopped doing all three things that path taught me: building, teaching and asking better questions.

For more than eight years I have built learning and assessment software, today as a senior full-stack developer working independently with BrainCore Sàrl. I teach programming and computer-science foundations, most recently as an external lecturer at Universiti Malaya. And as a PhD candidate, I research how machine-learning comparisons can stay honest.

In 2026 I founded PearlHelix, my own systems and advisory venture, to bring these strands to practical problems of learning and work.

Recognition

  • 1st

    in Malaysia, IEEEXtreme

    Twice, in 2021 and 2023, leading Universiti Malaya teams in IEEE’s global 24-hour programming competition. Six national placements from 2020 to 2025.

    The record

  • Gold

    LiTeR award, shared

    For physical games that help students grasp programming concepts, recognised by Universiti Malaya’s teaching-innovation programme.

    The work

  • Lead

    IEEE Malaysia Section

    Coordinating IEEEXtreme 20.0 across Malaysia in 2026: from competitor to the person helping the next teams compete.

    Leadership

02What I hold to

Understand deeply. Build thoughtfully.

  1. 01

    Understanding before answers.

    A correct answer without the reasoning is borrowed. I teach and build so that understanding becomes the learner’s own.

  2. 02

    Make the abstract tangible.

    Sorting, a variable, a 3D vessel, a rising river: an idea lands when you can act it out, see it or hold it.

  3. 03

    A result is only as honest as its comparison.

    Leakage, related samples and uneven budgets can flatter a model. I check the comparison before I trust the conclusion.

  4. 04

    Knowing when to stop is a skill.

    In a search, in research and in life, more effort is not always more progress. Spending attention wisely is its own kind of intelligence.

  5. 05

    Technology should leave people more capable.

    I use AI and automation where they strengthen human judgement, and protect room for independent practice where they would quietly replace it.

  6. 06

    Build for the person at the other end.

    Every permission rule, report and interface decides something about someone’s day. I keep that person in view.

03Selected work

Questions, turned into practice.

A few threads from my work in software, education and research. Each starts with a different question.

  1. 01Software engineering · BrainCoreConnecting the parts of a learning platformContent, permissions, activities, results and reports: the quiet work behind learning and assessment is making them fit together.
  2. 02Learning research · Universiti MalayaMaking an abstract idea something you can doPhysical games that let learners act out programming and computational-thinking concepts before returning to code.
  3. 03Applied machine learning · Universiti MalayaFinding patterns in cultural heritageClassification and retrieval experiments with 3D representations of Kendi artefacts.
  4. 04Collaborative research · ShonaBondhuAn early warning question, at system scaleA cloud-based research system for monitoring water levels and exploring flash-flood warning.
See all selected work

The question I am pursuing now

When should a search stop?

Automated machine learning can try hundreds of model configurations. Improvements often slow well before the budget is spent, yet the search keeps consuming time, compute and energy. My doctoral research studies search-level stopping, deciding when enough has been learned, and how to compare such methods without leakage flattering the result.

Illustrative chart: the best score found rises quickly over the first trials of a search, then flattens across the remaining budget.0.60.70.80.9050100150200Trials in the searchBest scoreStop
Stop after
40 of 200 trials
Compute saved
80%
Best score kept
96.7%
An illustration of the idea, not a research result.

Now · September 2026

What I am working on.

  • Doctoral research at Universiti Malaya on search-level stopping, evaluation validity and resource reporting in automated machine learning.
  • Building learning and assessment software, including ELIA, with BrainCore Sàrl.
  • Coordinating IEEEXtreme 20.0 across Malaysia as IEEE Malaysia Section Lead.
  • Shaping PearlHelix, my systems and advisory venture.

Open doors

Conversations I welcome.

  • Research collaboration on machine-learning evaluation, AutoML and reproducibility.
  • Learning technology, assessment and education systems.
  • Teaching, guest lectures and workshops in programming and computational thinking.
  • Software and advisory work, through PearlHelix.
  • Students and early-career engineers finding their way into computing.
Start a conversation

A conversation starts here

Let’s compare notes.

Whether it is a research idea, a system that needs untangling, a class to teach or a question you are still shaping, I would be glad to hear from you.

Save my contact Download CV
The printed card
ahsan@pearlhx.com