M. Reza Shahabian A.

AI Engineer · Agentic AI, LLMs & Multimodal ML · PhD

I work where AI meets the physical world — robots, wearables, and systems that help people live independently. These days that mostly means agentic AI: teams of specialist AI agents that can see, reason, and act. I recently built Linguagent, a live adaptive tutor run entirely by AI agents, and as a Research Fellow at the University of Hertfordshire I develop assistive AI for healthcare — from vision–language fall detection to wearable robots that sense what you intend to do. I also love teaching this stuff, from university modules to hands-on workshops on building your first AI agent.

Publishing as Mohamad Reza Shahabian Alashti.

View Projects Linguagent ↗ Publications GitHub

Projects & Research

eSAM — Exposure Sensing Animated Mannequin

Sep 2022 - Feb 2024

Human-sized robotic mannequin for chemical exposure testing with distributed sensors

RoboticsEmbedded SystemsSensor NetworksControl Systems

EEG & Facial Emotion Data Collection Platform

Feb-Oct 2025

Web-based platform for synchronized multimodal emotion data collection with EEG and facial recognition

EEGEmotion RecognitionVision TransformerWeb Platform

Data Study Group (DSG) — Advanced Manufacturing

Dec 2022

Collaborative data science research sprint addressing industrial manufacturing challenges

Data ScienceMachine LearningData AugmentationSynthetic Data

Linguagent — Adaptive AI Tutor for Language & Any Subject

2026

A personalized, adaptive learning tutor built as a team of specialist AI agents on Google ADK and Gemini, assessing learners against the CEFR standard (or a mastery scale for any subject) and tutoring them with feedback grounded in real rubrics.

AI AgentsGoogle ADKGeminiVertex AI

Hospital@Home / PRIME Study

2024-Present

AI-driven virtual ward system for ambient assistive care with multimodal patient monitoring

Healthcare AIMulti-View HARHRISensor Fusion

SWAG — Soft Wearable Assistive Garment

2024-Present

AI-driven intent-detection system using multi-sensor biomechanical data for wearable robotics

Deep LearningSensor FusionMeta-LearningTransformers

Multi-View Human Activity Recognition (RHM-HAR Series)

2020-Present

Comprehensive multi-view HAR research with benchmark datasets and advanced AI architectures

Human Activity RecognitionMulti-View VisionTransformerAgentic AI

Tutorials & Teaching

View all →

AWS VPC Foundations

An interactive introduction to Amazon Virtual Private Cloud (VPC): subnets, routing, gateways, and secure network design on AWS.

13 lessons

Information Security & ISO Compliance

Interactive learning suites for information security management and compliance: ISO/IEC 27001 (ISMS), ISO/IEC 27005 (risk management), ISO/IEC 27017 (cloud security), a hands-on risk-treatment activity, and UK cyber law & compliance.

5 lessons


Events & Workshops

From Prompt to Agent — Your First Intelligent AI System

📅 18 April 2026

📍 University of Hertfordshire, UK

A full-day hands-on workshop on Agentic AI at the University of Hertfordshire — designing, building, and deploying intelligent AI agents capable of autonomous decision-making.

ICSR 2026 — Special Session: Assistive Robotics and Exoskeletons for Human Augmentation and Wellbeing

📅 1-4 July 2026

📍 University of London, London, UK

Building interdisciplinary bridges through assistive technologies, social robots, and creative therapies


Publications

20+ publications · 67 citations · h-index 5 · i10-index 2 — Google Scholar

2026 ICRA 2026

A Modular, Wireless and Wearable Biosignal Acquisition Platform

Antonios Doukakis, Aikaterini Smyrli, Makis Livadas, et al.

2026 SSRN Preprint

A Deployment-Oriented Validation Framework for Cross-Day Wearable Activity Recognition Using EMG and IMU

Mohamad Reza Shahabian Alashti, Shadiya Alingal Meethal, Patrick Holthaus, et al.

2026 ICSR+Art 2026

EMG-Based Lower Limb Activity Recognition for Exoskeleton-Assisted and Unassisted Locomotion Using Deep Learning

Shadiya Alingal-Meethal, Mohamad Reza Shahabian Alashti, Martina Mosso, et al.

2026 Preprint (2026)

HAR-Agent: Multilingual Multimodal Activity Recognition via Knowledge-Distilled LLM Reasoning

Khashayar Ghamati, Mohamad Reza Shahabian Alashti, A. Fallahirahmatabadi, et al.

2026 ICSR+Art 2026

Pexformer: Robust Indoor Human Localisation via Patch-level Tokenisation and Semi-Permeable Attention

Baobing Zhang, Sehrish Rafique, Mohamad Reza Shahabian Alashti, et al.

2026 ICSR+Art 2026

From Pilot Data to Protocol: Sample-Size Guidance for Multimodal Intent Detection in Assistive Wearable Robotics

Mohamad Reza Shahabian Alashti, Shadiya Alingal-Meethal, Patrick Holthaus, et al.

2026 ICSR+Art 2026

Vision–Language Models for Fall Detection in Socially Assistive Robotics: Zero-Shot Prompting and Few-Shot Calibration

Mohamad Reza Shahabian Alashti, Khashayar Ghamati, Abolfazl Zaraki, et al.

2025 ICSR 2025

Towards Memory-Driven Agentic AI for Human Activity Recognition

Mohamad Reza Shahabian Alashti, Khashayar Ghamati, Hooman Samani, et al.

2024 BioRob 2024

Efficient Skeleton-based Human Activity Recognition in Ambient Assisted Living Scenarios with Multi-view CNN

Mohamad Reza Shahabian Alashti, Mohammad Bamorovat Abadi, Patrick Holthaus, et al.

2024 BioRob 2024

Robotic Vision and Multi-View Synergy: Action and Activity Recognition in Assisted Living Scenarios

Mohammad Bamorovat Abadi, Mohamad Reza Shahabian Alashti, Patrick Holthaus, et al.

2023 ACHI 2023

Lightweight Human Activity Recognition for Ambient Assisted Living

Mohamad Reza Shahabian Alashti, Mohammad Bamorovat Abadi, Patrick Holthaus, et al.

2023 ACHI 2023

RHM-HAR-SK: A multi-view dataset with skeleton data for Ambient Assisted Living Research

Mohamad Reza Shahabian Alashti, Mohammad Bamorovat Abadi, Patrick Holthaus, et al.

2023 ACHI 2023

RHM: Robot House Multi-view Human Activity Recognition Dataset

Mohammad Bamorovat Abadi, Mohamad Reza Shahabian Alashti, Patrick Holthaus, et al.

2022 Alan Turing Institute Report

Data augmentation and synthetic data generation for low-frequency and sparse data problems

Mohamad Reza Shahabian Alashti, Alan Turing Institute Team, AMRC Collaborators

2021 4th UKRAS21 Conference: Robotics at home Proceedings

Human activity recognition in RoboCup@ home: Inspiration from online benchmarks

Mohamad Reza Shahabian Alashti, Mohammad Bamorovat Abadi, Patrick Holthaus, et al.

2021 4th UKRAS21 Conference: Robotics at home Proceedings

Affordable Robot Mapping using Omnidirectional Vision

Mohammad Bamorovat Abadi, Mohamad Reza Shahabian Alashti, Patrick Holthaus, et al.

2018 6th RSI International Conference on Robotics and Mechatronics (IcRoM)

Automatic ROI Detection in Lumbar Spine MRI

Mohamad Reza Shahabian Alashti, Mohammad Reza Daliri, Behnam Jamei

2017 5th RSI International Conference on Robotics and Mechatronics (ICRoM)

FARAT1: An Upper Body Exoskeleton Robot

Farzad Cheraghpour, Farbod Farzad, Milad Shahbabai, et al.

2017 Artificial Intelligence and Robotics (IRANOPEN)

Mechanical Basic and Detailed Design for the Redundant Arm SAAM applied on a Domestic Service Robot

Farzad Cheraghpour Samavati, Majid Iranikhah, Parastoo Dastangoo, et al.


Professional Experience

Postdoctoral Research Fellow — Deep Learning & Robotics

University of Hertfordshire, UK

Aug 2024 - Present • Hatfield, UK

PythonPyTorchCUDAROS2HRNetYOLOv7

Visiting Lecturer

Various Institutions, London & Hertfordshire

Sep 2021 - Present • London & Hertfordshire, UK

TeachingAICloud ComputingAWSSoftware EngineeringResearch Methods

Senior Engineer & Project Lead

Arta Vision Ava Eng. Co., Tehran, Iran

Feb 2015 - Nov 2020 • Tehran, Iran

PythonDjangoDockerPostgreSQLGrafanaReactJS

Robotics Engineer

University of Hertfordshire, UK

Oct 2022 - Feb 2024 • Hatfield, UK

PythonArduinoRaspberry PiModbus RTUMaxonControl Systems

Mechatronics Designer

SYNTECH Technology & Innovation Center, Iran

Feb 2012 - Mar 2016 • Qazvin, Iran

MechatronicsRoboticsC++ROSLinuxEmbedded Systems

Education

PhD in Computer Science

University of Hertfordshire, UK

2020 - 2024

Thesis: Human Activity Recognition in Ambient Assisted Living Scenarios

MSc in Mechatronics Engineering

Azad University (IAU), Qazvin, Iran

2011 - 2015

Thesis: Intelligent System for Estimating Spinal Cord Injury using Medical Image Segmentation

BSc in Electronic Engineering

Hadaf University, Sary, Iran

2006 - 2011

Thesis: Design and Implementation of an Advanced Path-Following Robot


Skills & Expertise

AI & Machine Learning

Deep Learning, Computer Vision, Facial Emotion Recognition, Video-Based Human Activity Recognition, Pose Estimation, Sensor Fusion, Meta-Learning, Agentic AI, Vision Transformers (ViT), GANs

Programming & Development

Python, C, C++, JavaScript, ReactJS, Node.js, HTML/CSS

Frameworks & Libraries

PyTorch, TensorFlow, Keras, OpenCV, Scikit-Learn, NumPy, Pandas, Streamlit, QT, Docker, ROS1/ROS2

Robotics & Embedded Systems

ROS2, MoveIt, Navigation, Arduino, Raspberry Pi, Maxon Motor Control, Modbus RTU, Mechatronics Integration

Cloud & Infrastructure

Cloud Computing (AWS), Edge AI Deployment, Data Pipelines, Secure Cloud Storage, Scalable Web-Based Systems


About Me

I've been building machines since before I could name what I was doing. It started with electronics as a student — a path-following robot for my bachelor's — and grew into years with a RoboCup @Home team, where I led the electronics group and we won the national league five times. Somewhere along the way the question shifted from "can I make this move?" to "can I make this understand?" — and that question has carried me ever since.

Following it took me through industry — years as an engineer and project lead, shipping monitoring systems that ran quietly in data centres across the country — and then to the UK for a PhD in Computer Science, teaching machines to recognise what people are doing in their own homes. Today I'm a Research Fellow at the University of Hertfordshire, working on AI for assistive care: wearable robots that sense what you intend to do, and vision–language models that notice when something goes wrong. My side of the lab is where deep learning meets EMG signals, cameras, and people who need a bit of help to live independently.

Lately I've fallen for agentic AI — systems built not as one model but as a team of specialists that reason and act together. My playground for this is Linguagent, an adaptive AI tutor I designed and built end-to-end, live on the web and run entirely by agents. It's equal parts research lab and product, and it's taught me more about making AI reliable than any paper deadline ever has.

The other thing I can't stop doing is teaching. I lecture in AI, cloud computing, and software engineering; I chaired the special session on assistive robotics at ICSR 2026; and I ran "From Prompt to Agent", a full-day workshop where 55 people built their first AI agent with me. The tutorials on this site are part of the same habit — if I learn something worth knowing, I'd rather share it.

Full career details are on the resume page and LinkedIn. I'm based in the UK on a Global Talent visa.