AMAN^UG®

0+ Students
taught

0.80 ROC-AUC
risk model

0+ FTTH
layouts

SCROLL

01 — About

Systems thinking, from
the packet to the model.

I'm a Computer Systems Engineering graduate from Palestine Technical University – Kadoorie, seeking a computer engineer role where a solid academic foundation turns into working systems.

My work spans both halves of the stack. On the data side I build predictive pipelines in Python and Scikit-learn — feature engineering, model selection, and the unglamorous preprocessing that decides whether a model is honest. On the infrastructure side I've designed FTTH fiber layouts, RPA bots and REST integrations inside a national telecom.

I also teach. I run labs for 170+ students across programming, computer networks and Linux, and co-authored the manual for PTUK's first-ever Linux Lab — writing the curriculum I wish I'd had.

What ties it together: I like problems where the answer has to survive contact with real, messy, imbalanced data — or real cable in a real street.

Machine Learning Feature Engineering Data Pipelines Computer Networks FTTH / OSP Design RPA Automation Linux & Shell Teaching Technical Writing

NameAman Abuelayyan

RoleComputer Systems Engineer

DegreeB.Sc. Computer Systems Engineering

Average88.9%

GraduatedJuly 2025

LocationTulkarm, Palestine

FocusML · Data · Networks

StatusOpen to opportunities

02 — Track Record

Professional Experience

Teaching, machine learning and telecom infrastructure — three very different rooms, one way of working.

Assistant Research & Teaching Engineer

Oct 2025 — Present

Palestine Technical University – Kadoorie · Part-Time

Tulkarm, Palestine

Teaching assistance and lab support for five sections, 110+ students in Computer Programming Essentials (C++) — practical application, code debugging and reinforcing fundamentals.
Assisting the lead instructor with curriculum preparation, material development and grading lab assignments and quizzes.
Guiding 60+ students through the Computer Networks Lab: network topology design and packet analysis in Cisco Packet Tracer.
Co-authoring the manual for the inaugural Linux Lab — designing the first-ever curriculum on shell scripting, redirection, pipelining and system administration.

110+ C++ students 60+ networks students 1st Linux Lab curriculum

Data Science & Machine Learning Engineering Intern

Jan 2026 — Apr 2026

IBT Learning × GGateway · Part-Time

Remote, Palestine

Directed a team of 5 to architect a multi-stage Risk Scoring System for Brazil's largest department-store marketplace, combining XGBoost classification and regression to flag high-risk orders and estimate delay severity across 100k+ transactions.
Engineered a predictive pipeline in Python and Scikit-learn using geospatial Haversine distance and seller bottleneck metrics — 0.80 ROC-AUC on late-delivery detection despite an 11:1 class imbalance.
Owned the data lifecycle from raw relational tables to a 55-feature model-ready dataset, using unsupervised clustering to segment sellers into operationally meaningful risk profiles.

0.80 ROC-AUC 100k+ transactions 55 features 11:1 imbalance

IT & Network Infrastructure Engineering Intern

Jul 2024 — Aug 2024

Oman Telecommunications Company (OmanTel) · Full-Time

Muscat, Oman

Streamlined operational workflows with 5+ UiPath RPA bots, managing 10+ REST APIs and surfacing key performance metrics through Power BI dashboards.
Redesigned 15+ FTTH layouts for urban infrastructure, optimizing cable routing to cut material costs by 12%, while leading a GIS-OSP software migration for a 50-user technical team.
Resolved 20+ critical CRM bottlenecks and facilitated 10+ business requirement sessions to align cybersecurity protocols with operational infrastructure needs.

12% cost reduction 15+ FTTH layouts 5+ RPA bots 50-user migration

03 — Selected Work

Projects & Case Studies

Machine learning end to end — from raw relational tables to models that hold up on imbalanced, real-world data.

Browse all repositories

/ 01

Unsupervised Movie Review Clustering

An unsupervised NLP pipeline that sorts IMDB reviews into five thematic categories without a single label.

PythonK-MeansDBSCANWord2VecPCA

Clustered reviews with K-Means, DBSCAN and Agglomerative Clustering
Lifted Silhouette Score 22% by swapping TF-IDF for Word2Vec embeddings
Tuned feature extraction with PCA, validated cohesion via Calinski-Harabasz

View repository

/ 02

Marketplace Risk Scoring System

Multi-stage risk engine for Brazil's largest department-store marketplace — flags high-risk orders and estimates how late they'll be.

XGBoostScikit-learnPandasClustering

Classification plus regression across 100k+ transactions
0.80 ROC-AUC on late deliveries against an 11:1 class imbalance
Haversine geospatial distance and seller bottleneck features

/ 03

FTTH Network Redesign & RPA Suite

Fiber-to-the-home layouts re-routed for cost, plus the bots and dashboards that took the manual work out of the workflow.

UiPathREST APIsPower BIGIS-OSP

15+ urban FTTH layouts re-routed — 12% material cost reduction
5+ RPA bots and 10+ REST integrations replacing manual steps
GIS-OSP software migration delivered for a 50-user technical team

/ 04

PTUK Linux Lab — Curriculum & Manual

The university's first Linux Lab, written from scratch: a lab manual and exercise set for 60+ students learning the shell properly.

LinuxBashSys AdminTechnical Writing

Shell scripting, redirection and pipelining taught hands-on
System administration exercises built around real troubleshooting
Co-authored manual now the course's standard material

04 — Capabilities

Technical Skill Set

Depth in Python, data and systems, with working range across the web and BI stack.

Filled nodes are proficient · outlined are familiar

05 — Background

Education & Strengths

Sep 2020 – Jul 2025 · Tulkarm

B.Sc. in Computer Systems Engineering

Palestine Technical University — Kadoorie · Average 88.9%

Relevant coursework in Software Engineering, Artificial Intelligence and Machine Learning, alongside the hardware and networking core of a systems engineering degree.

Oct 2025 — Present

Teaching & Lab Development

PTUK — Programming, Networks and Linux labs

Running labs for 170+ students and co-authoring the curriculum for the university's first Linux Lab — turning coursework into material other people can teach from.

Ongoing

Applied Machine Learning

IBT Learning × GGateway · Credly

Continuing hands-on work in predictive modelling, feature engineering and evaluation on imbalanced real-world datasets, with verifiable badges on Credly.