Data & development
Python, data preparation and data systems. Turning messy inputs into usable foundations.
SABAH, MALAYSIA / PUBLIC ADMINISTRATION + TECHNOLOGY
Ahmad Hazwan bin Abinan
Data. AI.
Public service.
I bring public administration and data science together to build practical tools, understand complex problems, and support better decisions.
PUBLIC ADMINISTRATION / MSc DATA SCIENCE, UNIVERSITY OF EXETERSCROLL TO EXPLORE ↓
01 /ABOUT
An administrative perspective.
A technical foundation.
I currently serve the Sabah State Government. My MSc studies in Data Science at the University of Exeter connect that public service experience with hands-on work in analytics, machine learning and AI.
I am interested in what happens beyond a promising prototype: how a system performs in a real setting, how people use it, and how it can be operated responsibly.
University of ExeterMaster of Science in Data Science · 2025–2026
University of MalayaBachelor of Engineering · Manufacturing Engineering · 2008–2012
PUBLIC-SERVICE EXPERIENCE
Ministry of Local Government and Housing2020–2025
Sabah State Public Service Commission2018–2020
Analytics & machine learning
Model development, evaluation and site calibration, with a focus on real-world performance.
Strategy & governance
Evidence-based decisions, strategic and workforce analytics, data ethics and human oversight.
02 /SELECTED WORK
From research to reality.
FEATURED / MSc RESEARCH2026
Parking occupancy.
Beyond the benchmark.
Can a parking detection system stay useful when it leaves a benchmark dataset and runs on embedded hardware at a real site?
Read the case study
Operational dashboard linking model output, parking-space geometry and the live camera view.
Technical case study
Parking Occupancy Detection on Embedded Hardware: Accuracy, Efficiency, Transferability and Site Calibration
Problem & role
My MSc research examined parking occupancy detection across model families and datasets, with deployment on a Raspberry Pi 5. The work covered model development, evaluation, transferability and local calibration.
Technical approach
Classical machine learning, CNN/transformer models, object detection and vision-language models were explored. On-device deployment tested the practical relationship between accuracy and computational efficiency.
What I learned
Performance on benchmark data did not reliably transfer to a new site. Local calibration and evaluation across daytime and night-time conditions were central to making the system useful in practice.
Responsible operation
The operational design considered monitoring, retraining, performance-gated promotion and rollback. On-device inference and aggregate occupancy outputs support an approach focused on spaces, rather than identifying people.
Research project; this case study does not claim deployment as a Sabah government service. The linked project environments may require access.
03 /PUBLIC SERVICE
Useful possibilities for Sabah.
Areas where I would like to contribute, combining an understanding of administration with practical data and AI skills.
Find the right information
Search circulars and working documents with source references and access controls that respect each officer’s permissions.
Permission-aware search · citationsUnderstand service delivery
Use licensing, process and workforce data to understand delays, track meaningful indicators and support better resource decisions.
Process analytics · workforce insightsObserve facilities responsibly
Explore on-device AI for facilities and public assets, using aggregate insights and proportionate data collection.
Edge AI · privacy by designThese are proposals for exploration, subject to approval, data protection, risk assessment and human oversight.
04 /HOW I WORK
Build thoughtfully.
Check what matters.
I use Python and AI-assisted development tools, including Codex and Claude Code, for prototyping, automation, documentation and testing. Human review, data security and appropriate governance remain part of the work.
05 /CONTACT
Let’s start a
conversation.
For professional conversations, research and ideas worth exploring.