I build machine learning systems and the data infrastructure they run on. My research applies AI to healthcare, especially early cancer detection.
I care about the part of machine learning that comes after the notebook: making models reliable, measurable and useful to the people who depend on them.
I'm an AI/ML engineer and data scientist with ten years in software and data. I started in frontend and full-stack engineering, then moved through technical project management and data engineering into applied machine learning.
As a researcher I study how machine learning can support clinical decisions, especially cancer detection and healthcare analytics. I've published twelve peer-reviewed papers, including in IEEE venues. I'm also a tech entrepreneur, building companies that put AI and automation to work.
From writing frontend code to designing data platforms and machine learning systems.
I build AI and machine-learning capabilities end to end, from data modelling and storage through to the services that put models in front of users.
I found and build technology companies that use AI and automation to solve concrete business problems. I own the product direction and system architecture, and build the engineering teams.
I grew from junior engineer to data engineer, taking on more scope at each step: frontend, full-stack web, technical project management, then the full data-lake lifecycle.
My research uses machine learning to find disease earlier and help clinicians decide with better evidence.
Deep learning models that identify malignancy in clinical data and medical images, aimed at earlier, more accurate diagnosis.
Predictive models on patient and clinical data to support diagnosis, risk assessment and care decisions.
Hybrid networks that combine learned visual features with structured clinical signals.
Combining learners for robustness on noisy, imbalanced medical datasets.
12 peer-reviewed papers and 101 citations. The complete, up-to-date list is on Google Scholar.
All publicationsA security-first, standards-aligned framework for zero-touch, self-managing 6G networks, designed to run AI workloads reliably at the network edge.
Fuses CT imaging and clinical data from 2,000 patients in a hybrid CNN–DNN model. It reached 96.4% accuracy and an AUC of 0.982, beating standalone CNNs and classical machine learning.
A review of how growing data volume, variety and complexity are reshaping database systems, and the architectures emerging to handle them at scale.
Where I follow current research in AI and data science and meet the people doing it.
Princeton University
Dhaka, Bangladesh · attended online
Monroe University
BRAC University
muhammodrayhan@gmail.comCopied to clipboard