About Me
Hi, I'm Dr. Madhurananda (Madhu) Pahar, an Applied AI Scientist and Senior Research Fellow working at the intersection of machine learning, speech AI, multimodal AI, and healthcare. I enjoy building AI systems that solve real-world problems—from early dementia detection to full-stack AI applications.
India → Scotland → South Africa → USA → England: My journey across continents in pursuit of knowledge and research.
Early Life & Education
I was born and raised in a historic village of Dhanyakuria, near Kolkata, India. After doing my primary school there, my family moved to Madhyamgram, closer to Kolkata, where I did my secondary school.
After completing my BSc in Mathematics at Presidency College, under the University of Calcutta, I moved to the University of Stirling, Scotland, where I completed my MSc in Computing for Financial Markets and PhD in Computational Neuroscience.
Research Journey
I moved to Stellenbosch University in South Africa, where I did my first postdoc on identifying tuberculosis (TB) and COVID-19 from vocal audio (cough, breath, and speech), and long-term monitoring of patients using accelerometer (sensor) data, under Professor Thomas Niesler. During this time, I developed AI systems for detecting tuberculosis (TB) and COVID-19 from cough, breath, and speech recordings, while also working on the long-term monitoring of patients using wearable accelerometer data.
My next journey took me to the University of Nevada, Las Vegas (UNLV), where I was a research scholar at the QGen Lab, analysing long-read third-generation Oxford Nanopore Technology (ONT) Sequencing using deep learning, under Dr. Q. Chris Liu. There, I applied deep learning to third-generation sequencing data, expanding my experience beyond speech AI into genomics.
I am currently based in Sheffield, working as a Senior Research Fellow and Applied AI Scientist at the University of Sheffield, focusing on identifying cognitive decline such as dementia and mild cognitive impairment (MCI) using speech audio, under Professor Heidi Christensen.
In this role, I have also mastered an end-to-end ML pipeline using multimodal AI on audio, text, and image (facial cues from captured video) from the CognoSpeakplatform.
Side Projects & Interests
Beyond research, I enjoy solving real-world problems through AI. For example, I developed a map distribution system for the local congregation as a full-stack web developer.
Presenting research at the Pint of Science festival
I also enjoy travelling, exploring different cultures, teaching the general public about AI, and cooking.
I'm passionate about building AI that has a measurable real-world impact. Whether through healthcare research, open-source software, or modern AI applications, I enjoy transforming complex ideas into practical systems that people can use.