Profile banner
Ayman Elhalwagy

Ayman Elhalwagy

AI researcher and founder building causal decision intelligence

ayman
England
Joined March 2025
Favicon imageaymanelhalwagy.com
Favicon imageorcid.org
Favicon imagescholar.google.com
Favicon imagetwitter.com
Favicon imagelinkedin.com

Network

312 connections
💰
Deep Tech Investors
💡
Startup Operators Advisors
📊
AI Data Leaders
⚖️
Startup Legal Finance
📈
...
🌐
...
🧠
RootCause.ai Team

Summary

Researcher in energy‑efficient, generalisable deep learning for industrial condition monitoring — Ayman's PhD work focuses on architectures and data‑fusion methods that improve cross‑dataset generalisation while reducing compute (principles of Green AI). His thesis and related papers introduce a Multi‑Channel LSTM‑Capsule Autoencoder, a Dataset Fusion algorithm and an Order Domain Transformer aimed at making ML for industry more data‑efficient and deployable. ac+2
Operator of AI startups that translate research into commercial products — He co‑founded Numvio (2022) to deliver AI R&D and later founded Perceptura (rebranded RootCause) to productise causal decision intelligence for enterprises, positioning himself at the intersection of deep research and practical, revenue‑driving AI applications. aymanelhalwagy+2
Proven publication and citation record in applied ML and condition monitoring — Multiple peer‑reviewed articles, conference papers and citations (Google Scholar profile lists multiple journal and conference publications), demonstrating an active academic contribution alongside entrepreneurial work. google+2
Bridges academic impact and real‑world deployment — His work has been trialled by external organisations (personal site and bios reference trials with NASA and mining portfolios) and has earned institutional recognition (Brunel Research Impact Award referenced in bios), indicating ability to take research beyond publications into operational impact. aymanelhalwagy+2

Work

Education

Projects

Writing

Green AI for industry 4.0: Energy‑efficient generalised deep learning approaches to induction motor condition monitoring

January 1, 2025

PhD thesis presenting LSTM‑Capsule autoencoder architectures, a Dataset Fusion algorithm and an Order Domain Transformer to improve cross‑motor generalisation and reduce computational cost for industrial condition monitoring.

Favicon imagebura.brunel.ac.uk

A Dataset Fusion Algorithm for Generalised Anomaly Detection in Homogeneous Periodic Time Series Datasets

January 1, 2023

Paper describing the Dataset Fusion method to combine periodic time‑series datasets for improved generalisation of anomaly detection models (arXiv / IEEE Access publication venues noted).

Favicon imagearxiv.org

Multi‑Channel LSTM‑Capsule Autoencoder Network for Anomaly Detection on Multivariate Data

January 1, 2022

Journal article describing an LSTM‑Capsule autoencoder for multivariate anomaly detection, demonstrating resilience to overfitting and improved training efficiency for condition monitoring tasks.

Favicon imagemdpi.com