Ubada El Joulani

Ubada El Joulani1

AI Engineer & Researcher

1PhD Researcher, Electrical and Electronic Engineering, Brunel University of London.

PhD researcher building AI products that solve real-world problems. From energy forecasting to condition monitoring to the development of literature review tools, turning research into impact.

View projects Google Scholar LinkedIn Get in touch
Portrait of Ubada El Joulani
Fig. 0. The author.
Currently
Doctoral Researcher, Brunel University of London
Research
Fault diagnosis and data-efficient machine learning
Building
PaperSynapse, World Cup Predictor, FlipCar UK
I. About

Building intelligent systems.

Abstract—I’m a PhD researcher in Electrical and Electronic Engineering at Brunel University of London, developing AI-enabled condition monitoring for robotics in fusion remote maintenance. My work focuses on fault diagnosis and data-efficient machine learning: understanding how operating conditions distort learned fault features, and reducing computational demand without sacrificing predictive performance.

I graduated with a First Class BEng (Hons) in Computer Systems Engineering, including a one-year placement at the Brunel Innovation Centre in Cambridge, where I worked on applied AI projects spanning terahertz imaging, supply chain optimisation, and contraband detection.

Beyond research, I build products. I created PaperSynapse, an AI-powered tool that automates data extraction and analysis for systematic literature reviews; World Cup Predictor, which forecasts match outcomes from historical results, team and player statistics, and tournament probabilities; and FlipCar UK, a vehicle analysis platform that combines MOT history, market valuations, and deal scoring to help UK buyers avoid overpaying. I’m driven by the belief that technology should turn data into better decisions.

Index Terms: fault diagnosis, condition monitoring, predictive maintenance, data-efficient machine learning, time-series analysis, AI products.

Table I
At a Glance

Live productsYears in AI/MLLanguages spokenClass honours
54+41st
0.0 Nm

Nearest-centroid accuracy (toy model): 100%

Fig. 1. Why fault diagnosis models struggle under load (interactive; drag the slider). An illustrative, simulated feature space of motor-current signals. With no load, healthy and faulty signals form distinct clusters. As load rises, operating conditions dominate the learned features: clusters spread (dashed circles) and overlap, and classification suffers. My PhD research measured this effect on KAIST experiments, where 1D-CNN accuracy fell from 100% at 0 Nm to 53.19% at 4 Nm. See also [1].
II. Projects

Things I’ve built.

  1. PaperSynapse

    AI-Powered Systematic Literature Review

    Automates extraction and analysis of data from academic papers using AI. Imports from Scopus, PubMed, IEEE and more via CSV/RIS, with Claude-powered abstract extraction, screening, label normalisation, interactive analytics, AI chat with dataset-querying tools, and customisable fields that cut review time from 40+ hours to minutes.

    Keywords: AI-powered data extraction, analytics, academic research

    papersynapse.com

  2. World Cup Predictor

    AI-Powered World Cup Match Predictions

    Predicts World Cup match outcomes using machine learning models trained on historical tournament data, team rankings, and player statistics. Get data-driven insights on group stage results and knockout round probabilities.

    Keywords: machine learning, sports analytics, predictions, football

    predictorworldcup.com

  3. FlipCar UK

    Know What It’s Worth, Before You Buy

    A vehicle analysis platform for UK car buyers and flippers. Combines DVLA & MOT history, live market valuations from Auto Trader, Motors and Gumtree, mileage-based fault prediction, and an automated Deal Score so you never overpay for a used car.

    Keywords: DVLA and MOT data, market valuation, deal score, car flipping

    flipcaruk.com

Also built

  • Khutba Labs(contributed)

    Developed the main website and contributed to Lisaan AI, which translates live khutbas and Islamic lectures into 30+ languages.

    khutbalabs.com

  • The Safety Savvy

    Website for an accredited UK first aid training provider, showcasing six FAA-accredited, Ofqual/RQF-regulated courses.

    thesafetysavvy.com
III. Experience

Where I’ve worked.

  1. Oct 2023 – PresentUxbridge, UK

    Doctoral Researcher

    Brunel University of London

    • Developing AI algorithms for subsystem condition monitoring, predictive maintenance and fault classification in robotics for fusion reactors and other challenging environments
    • Analysing motor-current time series to investigate how load, speed and temperature influence fault-related features and model accuracy
    • Investigating data selection and unsupervised learning to reduce processing requirements and associated CO2 emissions while maintaining predictive performance
    • Established load-sensitivity benchmarks, showing 1D-CNN accuracy falling from 100% at 0 Nm to 53.19% at 4 Nm on KAIST experiments
    • Validated load-dominance patterns on the KAIST and Paderborn datasets using feature-space analysis, and linked cluster spread and diameter to fault-cluster purity as evidence for unsupervised stopping criteria
    • Led methodology, software implementation, validation and manuscript preparation for a first-author Applied Sciences article

  2. Sep 2021 – Sep 2022Cambridge, UK

    Project Technical Assistant (Placement)

    Brunel Innovation Centre

    • SCORE (Supply Chain Optimisation for Demand Response Efficiency): improved inventory visibility and production downtime forecasting by implementing tracking sensors and building demand-forecasting and inventory-change models with Python, pandas, MATLAB and Excel
    • ATTIC (Automated Terahertz Imaging of Composites and Tooling Profiling): enhanced composite-machining and parts-inspection accuracy with AI software for laser-profiler monitoring and terahertz inspection, using image classification, segmentation and PCA
    • Developed a neural-network model to identify contraband in prisons from safe, continuous passive terahertz imaging
    • Conducted applied AI research with multidisciplinary teams using Python and TensorFlow, and directed market research to inform engineering projects

  3. Jun 2021 – Sep 2021High Wycombe, UK

    Software Developer

    Elodiz Ltd

    • Developed software for laboratory equipment using C++
    • R&D with Python for data analysis to investigate software development approaches
    • Built an Arduino-controlled platform with stepper motors for precise X, Y and Z positioning in spectroscopic sample analysis
    • Delivered functional physical prototypes designed in Fusion 360 and fabricated through 3D printing

  4. May 2017 – Sep 2017Bergamo, Italy

    Web Developer

    Plat1 s.r.l.

    • Built responsive websites and web features for small businesses using HTML5 and CSS
    • Managed domain registration, multimedia integration, Google Analytics, and SSL implementation
    • Digitised invoices and maintained billing records in InvoiceX in line with Italian invoicing requirements
IV. Publications

Research output.

  1. [1]

    U. El Joulani, T. Kalganova, and S. Pamela, “Unsupervised Feature Space Analysis for Robust Motor Fault Diagnosis Under Varying Operating Conditions,” Applied Sciences, vol. 16, no. 4, p. 1780, 2026.

    doi:10.3390/app16041780
  2. [2]

    U. El Joulani, T. Kalganova, S.-A. Mitoulis, and S. Argyroudis, “AI and Remote Sensing for Resilient and Sustainable Built Environments: A Review of Current Methods, Open Data and Future Directions,” arXiv preprint, 2025; updated preprint on Research Square, 2026.

    arXiv:2507.01547 doi:10.21203/rs.3.rs-10272145/v1

Full list on Google Scholar

V. Education

Academic background.

  1. Oct 2023 – Present

    PhD Electrical and Electronic Engineering

    Brunel University of London

    Thesis: AI-Enabled Subsystem-Level Condition Monitoring in Fusion Remote Maintenance

  2. Sep 2019 – Jun 2023

    BEng (Hons) Computer Systems Engineering

    Brunel University London

    First Class Honours with Professional Development. Final Year Project: Smart Homes - Intelligent Data Collection and Processing (home-energy management algorithms for residential electricity-demand forecasting and non-intrusive appliance load monitoring using edge computing, with LSTM, ARIMA, moving-average and exponential-smoothing methods, and Dynamic Time Warping for unequal-length time series). Also developed Java banking software applying OOP principles.

  3. Sep 2013 – Jun 2019

    Science & Maths High School with Applied Sciences

    Istituto Maironi Da Ponte & Leonardo Da Vinci, Bergamo, Italy

    Erasmus+ programme in France and Portugal

VI. Leadership & Volunteering

Giving back.

  1. 2025

    Postgraduate Research Symposium Co-Chair

    Brunel University of London

  2. 2020 – 2022

    Peer Assisted Learning Leader & Senior PAL Leader

    Brunel University London

  3. 2020 – 2022

    Brunel Buddy Mentor & Student Services Ambassador

    Brunel University London

  4. 2015 – 2019

    Volunteer

    Islamic Relief Italy

VII. Skills

My toolkit.

Languages & frameworks
Python, Java, C++, C, HTML5, CSS, SQL, Assembly
AI & machine learning
TensorFlow, PyTorch, deep learning, neural networks, CNNs, LSTM/RNN, unsupervised learning, computer vision, NLP, LLMs
Methods
Image classification, image segmentation, PCA, t-SNE, K-means, hierarchical clustering, ARIMA, exponential smoothing, dynamic time warping, model evaluation
Tools & platforms
MATLAB, Simulink, SPSS, pandas, NumPy, SciPy, scikit-learn, Matplotlib, Seaborn, Fusion 360, Excel, Arduino, MPLAB, PIC microcontrollers
Domains
Fault diagnosis, condition monitoring, predictive maintenance, time-series analysis, data efficiency, edge computing, IoT, embedded systems, data analysis, computer architecture, network security, AI prototyping, research communication
VIII. Languages

I speak four.

Table II
Language Proficiency (CEFR)

Language A1A2B1B2C1C2 Level
ItalianNative ●●●●●● Native (C2)
ArabicNative ●●●●●● Native (C2)
EnglishAdvanced ●●●●●○ Advanced (C1)
FrenchElementary ●●○○○○ Elementary (A2)
IX. Contact

Let’s connect.

Open to collaborations, research opportunities, and interesting conversations.

ubadaej@hotmail.com
LinkedIn Google Scholar +44 7305 963533