AI Data Scientist

Location
Work from home in Ireland.
Eligibility to work
Unfortunately we cannot offer Irish work permits or Visa sponsorship.
About this Role
As a Senior AI Data Scientist in Research Pathology at Deciphex, you will lead advanced data-science research supporting the development and validation of AI for toxicologic pathology and translational research. Working at the interface of histopathology, computational pathology, and machine learning, you will translate complex scientific questions into rigorous datasets, experiments, benchmarking frameworks, and qualified analytical evidence.
Working closely with pathologists, AI researchers, software engineers, and product teams, you will ensure that model development is grounded in real tissue morphology, diagnostic reasoning, and unmet research needs. You will also support translational research activities, including biomarker discovery, IHC quantification, and tissue-based endpoint characterisation, and help convert exploratory research into reproducible methods, scientific publications, evidence packages, and future product capabilities.
You will play a leading role in pathology foundation-model research, designing training experiments, developing practice-relevant benchmarks, and translating exploratory research into reproducible methods and downstream pathology applications.
Key Responsibilities
Curate, characterise, and analyse large preclinical and translational pathology datasets, including whole-slide images, structured study data, annotations, and associated metadata
Lead the design and execution of advanced data-science and machine-learning research across toxicologic pathology and translational research applications
Design and conduct pathology foundation-model training experiments, including data-curation studies, training-recipe ablations, fine-tuning, and evaluation of learned representations for downstream pathology applications
Validate downstream pathology AI models through performance characterisation, confounder analysis, and evidence packages suitable for regulatory scrutiny
Build benchmarking datasets and practice-relevant evaluation tasks for foundation and downstream models, including non-clinical benchmarks grounded in whole-slide pathology use where no public standard exists
Develop and standardise data-processing pipelines and validation routines that support model development, benchmarking, and deployment
Support translational research applications, including biomarker discovery, IHC quantification, and tissue-based endpoint characterisation
Work directly with pathologists to ground datasets and model outputs in real morphological findings, lesion terminology, and diagnostic reasoning
Collaborate with the AI, software, and product teams to align research outputs with Deciphex's scientific and product roadmap
Required Skills and Experience
PhD in data science, bioinformatics, computational biology, biomedical science, statistics, computer science, or a related quantitative field (equivalent research experience may also be considered)
Strong proficiency in Python and relevant machine-learning and scientific-computing frameworks, such as PyTorch and scikit-learn.
Experience curating, integrating, and quality-controlling large imaging datasets and associated structured metadata
Demonstrable experience applying data science and machine learning to digital pathology, biomedical imaging, or complex biomedical datasets
Strong grounding in applied statistics and model validation, including experimental design and selection of clinically or operationally meaningful performance measures
Experience designing and conducting deep-learning experiments, including model training, evaluation, ablation studies, and systematic comparison of modelling approaches
Ability to translate scientific questions into well-defined datasets, analytical plans, validation strategies, and clear evidence-based conclusions
Experience developing reproducible analytical workflows using appropriate practices for version control, testing, environment and configuration management, provenance, and technical documentation
Ability to read and critically interpret scientific publications, technical standards, and relevant regulatory guidance
Excellent written and verbal communication in English, including the ability to produce clear technical and scientific documentation for specialist and non-specialist audiences
Demonstrated ability to work effectively across multidisciplinary and multi-organisation teams, managing competing scientific and delivery priorities
Desirable Skills and Experience
Working knowledge of histopathology, including tissue morphology, pathological findings and terminology, and the ability to discuss datasets and model outputs fluently with pathologists
Understanding of preclinical toxicology study design, treatment and control groups, endpoints, and histopathology workflows
Experience with whole-slide imaging, digital pathology platforms, image formats, annotation workflows, and large-scale image-data management
Experience with pathology foundation models, self-supervised learning, Vision Transformers, representation learning, model fine-tuning, or knowledge distillation
Experience training and evaluating deep-learning models using GPU, HPC, cloud, or distributed-computing infrastructure
Experience with translational pathology applications, such as biomarker discovery, IHC or multiplex-image analysis, tissue segmentation, and tissue-based endpoint characterisation
Experience collaborating with software engineers to translate research models and analytical workflows into maintainable, tested, and product-ready implementations
A record of peer-reviewed publication, conference presentation, or substantive delivery of applied AI research
About the Company
Through the work that we do, the team at Deciphex helps pharma to accelerate the process of essential drug development and helps patients to get timely and accurate diagnosis.

Employment type
Full-Time,Permanent

Work arrangement
Yes