Muhammad Junaid Ali

Data Scientist — Healthcare AI Developer

PhD Scholar — IRIMAS (UHA)
Medical Imaging • AutoML • NAS
Contact GitHub

About

I am a third-year PhD research scholar at the IRIMAS Lab, Université de Haute-Alsace (UHA), Mulhouse, France. Previously, I worked at Gomal University and Virtual University of Pakistan as a Tutor/Instructor teaching multiple computer science courses.

Before that, I was a research assistant at the Medical Imaging and Diagnostics (MID) Lab, where I proposed approaches for brain tumor segmentation, survival prediction, and breast pectoral muscle segmentation.

Focus
  • AutoML & Neural Architecture Search (2D/3D)
  • Medical image classification & segmentation
  • Robust NAS and multi-objective optimization

Work Experience

Ph.D. Research Scholar — UHA | Dec 2021 – Dec 2024

  • Designed AutoML approaches for medical image analysis.
  • Proposed NAS methods for 2D/3D segmentation and classification.
  • Developed robust NAS against adversarial attacks.
  • Formulated multi-objective NAS for lightweight architectures.

Technical Skills

Data Science
A/B testing, optimization, big data pipeline (cleansing, wrangling, visualization, modeling, interpretation), AutoML, Image Processing
Medical Imaging
MONAI, nnUNet, nibabel, pydicom
Web
Flask
Programming
Python (Pandas, scikit-learn, pytest, TensorFlow, PyTorch, SciPy, NLTK, Gensim), SQL, R, C++, Java
Cloud & Big Data
AWS (SageMaker, ECR, EMR, S3, Redshift), Spark, Databricks, Airflow

Projects

Evolutionary NAS for 2D/3D Medical Image Classification

Proposed an evolutionary NAS approach for medical image classification and a comparative study of multiple metaheuristics.

More details

Robust NAS Using Differential Evolution for Medical Images

DE-based NAS with attention-rich search space across convolution and pooling operations.

More details

Tashkhees (Breast Cancer AI Diagnostic App)

Integrated segmentation AI modules and built a Flask frontend for clinical workflow.

More details

CNNs via Surrogate-Assisted GA

Search of CNN architectures for medical image classification using surrogate models.

More details

Brain Tumor Segmentation & Survival Prediction

DL pipeline for segmentation and survival prediction; ranked in BraTS 2020.

More details · Challenge Ranking

Hardware-Aware NAS (NSGA-II)

Multi-objective NAS for lightweight medical imaging architectures. Under review.

Training-Free U-Net for Retinal Vessel Segmentation

NAS for U-Net shaped architectures. Under review.

More details

Selected Publications

Review of AutoML Optimization Techniques for Medical Image Applications

Computerized Medical Imaging and Graphics

Multi-task learning architecture for brain tumor detection and segmentation in MRI images

Journal of Electronic Imaging 31(5), 2022

Multi-level Kronecker CNN (ML-KCNN) for Gliomas Segmentation from Multi-Modal MRI

Journal of Digital Imaging

Enhancing breast pectoral muscle segmentation using skip connections in FCN

Int J Imaging Syst Technol, 2020

U-Net Based Glioblastoma Segmentation with OS Prediction

Intl. Symposium on Intelligent Computing Systems, 2020

Designing CNNs with Surrogate-Assisted GA for Medical Image Classification

GECCO 2023

Evolutionary NAS for 2D/3D Medical Image Classification

ICCS 2024

Robust NAS using Differential Evolution for Medical Images

EVOAPPS 2024

Full list available on Google Scholar.

Education

Ph.D. in Computer Science
Université de Haute Alsace — 2021–Present
MS Computer Science
COMSATS University Islamabad — 2018–2020
BS Computer Science
National University of Computer and Emerging Sciences — 2013–2017

Awards

BraTS 2020 Overall Survival Prediction — 3rd Place
MICCAI Society — Intel monetary award (2020)
BRATS 2020 Challenge
100% Ph.D. Scholarship
University of Upper Alsace (2021–2024)
100% BS Scholarship
National University of Computer and Emerging Sciences (2013–2017)

Languages

English
Fluent (PhD studies in English)
French
A1
Urdu
National Language
Punjabi
Native

Contact

The best way to reach me is email. I'm open to collaborations and applied AI projects in medical imaging.

You can connect this form to Formspree/Basin or a serverless function.