Lecturer & Researcher · Politeknik Negeri Samarinda

Fajerin
Biabdillah.

I build machine-learning systems that have to work on cheap hardware, unreliable networks and real users — from smart meters and river sensors in East Kalimantan to players, learners and immersive media.

IoT & edge MLTime-series & behavioural dataHCI, games & XRBlockchain systems
Portrait of Fajerin Biabdillah
Open to PhD positions · 2027
Scroll
01

About

I am a lecturer in the Department of Information Technology at Politeknik Negeri Samarinda, East Kalimantan, Indonesia, where I teach human–computer interaction, mobile and Android-for-IoT programming, operating systems, computer architecture and data mining.

I hold a Master of Computer Science from Universitas Brawijaya, where I was part of the Media, Game and Mobile Technology research group, and a Bachelor in Informatics Engineering Education from Universitas Negeri Malang.

My research asks a practical question: how do we make machine learning trustworthy when the data are noisy, the hardware is cheap and the conditions keep changing? I answer it with field deployments, careful evaluation and reproducible code.

0papers published or accepted, 2025–26
0manuscripts under review or in revision
0studies in my reproducible systematic review
0players profiled with deep clustering
02

Research focus

IoT & edge machine learning

Fault and anomaly detection on low-cost sensing hardware, evaluated under realistic false-alarm budgets and on sites the model has never seen.

Behavioural & time-series data

Sequence and multimodal models of how people play, learn and look — from game telemetry and eye tracking to market and news signals.

HCI, games & immersive media

Human-centred design of interactive and educational experiences, evaluated with users — and the next step: experiences that adapt to them.

03

Selected projects

2026Major revision · BEEI

Smart-meter fault detection under false-alarm budgets

A re-benchmark of fault detection on public 1 Hz ESP32 + PZEM-004T metering data (three stations, 1.8 million samples). The benchmark's own event-F1 protocol turned out to be degenerate — an always-on detector scores a perfect 100 — so I replaced it with a false-alarm-budgeted, leakage-safe protocol. Scale-invariant features nearly doubled AUPRC; transfer to an unseen station remained the binding constraint.

0.350 → 0.642mean AUPRC
74%events caught at 1% alarm budget
1 dayof labels closes the new-site gap
LightGBMtime seriesleave-one-site-outwith A. Triyono
2025–26Field deployments

IoT in the field, East Kalimantan

A Raspberry Pi salinity logger in the Sungai Wain conservation forest, reachable over 4G only through reverse tunnelling behind carrier-grade NAT; low-cost RTL-SDR AIS receivers for forecasting tugboat overspeed on the Mahakam river.

99.4%tunnel uptime
0data lost in 72 h
Raspberry Pi4G/CGNATAISICOMIT 2026
2025–26Presented · submitted

Modelling player behaviour

Multi-modal deep clustering of Dota 2 players from performance, team and hero-selection features, and attention-based sequence models transferred from Dota 2 to Mobile Legends.

5,000players profiled
4behavioural profiles
deep clusteringattention-LSTMtransfer learning
2017–21Published · JITeCS

Go Story — learning history by podcast

My Master's thesis: a gamified mobile learning app built with human-centred design and A/B testing, evaluated in a controlled experiment with 61 students.

86.84%usability
67.15%mean learning gain
HCDADDIEcontrolled experiment
2025Published · Atlantis Press

Zakat Chain & blockchain systems

Smart contracts and a token for transparent zakat management, evaluated with users; and BAWAS, a Hyperledger Fabric design for tamper-evident website-integrity auditing.

89.9%PSSUQ satisfaction
SolidityHyperledger Fabricusability
2026In revision · JAIC

A reproducible review of blockchain & AI in finance

A PRISMA-2020 systematic review built entirely as open Python scripts, from search to coding, with a framing classifier validated on an independent sample.

1,751studies included
PRISMA 2020Pythonopen pipeline
2026Best Presenter · PI

Sensing prototypes: wildlife and home gardens

An IoT and AI prototype for monitoring wildlife biodiversity in tropical forest (Best Presenter, Wehea-Kelay biodiversity symposium), and, as principal investigator of a 2026 research grant, an ESP32 + Flutter system with a web dashboard for monitoring home-grown vegetables.

IoTconservationESP32Flutter
04

Publications

05

Experience & education

Experience

  • 2024 — nowLecturer (Asisten Ahli)Department of Information Technology, Politeknik Negeri Samarinda
  • 2026Programme coordinatorZJIET Training Camp on Robotics Technology for Indonesian Teachers, Zhejiang, China
  • 2024 — 2025R&D Manager (part-time)PT Artavizta Karya Teknologi, Kalimantan branch

Education

  • 2017 — 2021M.Kom., Master of Computer ScienceUniversitas Brawijaya · Media, Game and Mobile Technology group · thesis supervised by Dr.Eng. Herman Tolle and Dr.Eng. Fitra A. Bachtiar
  • 2012 — 2017S.Pd., Informatics Engineering EducationUniversitas Negeri Malang

Teaching

Human–Computer InteractionMobile ProgrammingAndroid Programming for IoTOperating SystemsComputer ArchitectureData MiningSystems Analysis & Design

Toolbox

Python · PyTorchscikit-learn · LightGBMTransformers · LSTM · GNNESP32 · Raspberry PiFlutter · AndroidSolidity · HyperledgerGit · LaTeX
Best PresenterJoint Symposium on Biodiversity Wehea-Kelay, 2026
Principal InvestigatorPDP research grant, 2026
Registered copyrightEC002026100598 · barcode stock-taking app, 2026

Let's build something
that works in the field.

I am open to doctoral positions from 2027 and to research collaborations on IoT, machine learning for behavioural data, and interactive media.