Data Scientist – Cancer Diagnostics

About the role

MiRXES is seeking a Data Scientist to lead data‑science, bioinformatics and machine‑learning projects that underpin the development of cancer diagnostic products. The role will analyse in‑house clinical and molecular datasets to discover biomarkers, build diagnostic models and provide data‑driven insights that shape assay design, product development and clinical study strategies.

Key responsibilities

• Lead data analysis and machine‑learning projects for cancer biomarker discovery using multi‑omics and clinical data.
• Identify novel DNA, RNA and ncRNA biomarkers and develop multi‑marker diagnostic models.
• Design and implement analytics workflows covering data processing, quality control, model training, validation and performance evaluation.
• Support AI capabilities, including medical image analysis (e.g., CT, X‑ray) where relevant.
• Collaborate with clinical, assay development, product and commercial teams to translate insights into product strategy.
• Mentor junior data scientists and bioinformaticians, fostering technical growth and knowledge transfer.
• Enhance data literacy across R&D and promote data‑driven decision‑making.

Required profile

• Master’s degree or PhD in Data Science, Computer Science, Bioinformatics, Computational Biology, Biomedical Engineering, Statistics or a related quantitative discipline.
• 2–5 years of experience developing machine‑learning or deep‑learning models for biomedical, healthcare, genomics or life‑science applications (or equivalent postgraduate research experience).
• Proven experience working with high‑dimensional biological or clinical datasets, such as blood biomarkers, genomics, sequencing data or biomedical images.
• Hands‑on experience taking ML projects through the full workflow: preprocessing, feature engineering, model development, validation, performance evaluation and interpretation.

Required skills

• Python for scientific computing and machine learning.
• PyTorch for deep‑learning model development.
• scikit‑learn and conventional ML techniques (logistic regression, random forests, gradient boosting, SVM, clustering, dimensionality reduction, feature‑selection).
• Deep‑learning architectures: CNNs, transformers, attention‑based models, autoencoders.
• Biomedical image analysis (pre‑processing, augmentation, segmentation, classification).
• Next‑generation sequencing data analysis (FASTQ, BAM/CRAM, VCF formats).
• Model validation best practices (train/validation/test splits, cross‑validation, hyper‑parameter optimisation, handling class imbalance).
• Performance metrics for diagnostic models (ROC‑AUC, sensitivity, specificity, precision/PPV, NPV, calibration).