Junior Data Scincetist

Job Requirements

Role:

To contribute to the end-to-end development and product lifecycle management of artificial intelligence and machine learning (AI/ML) systems as part of an R&D organization. The Junior Data Scientist supports the definition of intended algorithm performance, the development and curation of representative data and annotations, the development of AI/ML algorithms, and the generation of scientific evidence demonstrating stand-alone algorithm performance.

The candidate is expected to:

- Apply scientific, statistical, machine learning, and software engineering methods to develop and evaluate AI/ML algorithm components for defined product use cases.
- Work with experienced Data Scientists, product management, clinical or domain experts, software developers, and quality and regulatory functions to translate customer needs and the intended use into measurable algorithm performance targets.
- Support the definition, sourcing, curation, documentation, and maintenance of representative data cohorts and annotations required for training, tuning, and evaluation.
- Plan and execute reproducible scientific evaluations that provide objective evidence of algorithm performance, limitations, and suitability for the intended use.
- Contribute the Data Science work products and evidence required for integration of AI/ML algorithm components into products and for their verification, validation, regulatory release, and lifecycle monitoring.

What are my responsibilities?

As a Junior Data Scientist, you are required to:

- Support the finalization of product and algorithm concepts by translating customer needs, intended use, clinical or technical requirements, and applicable regulatory considerations into measurable state-of-the-art performance targets.
- Contribute to the definition of the AI/ML development and scientific evaluation strategy for assigned algorithm components under the guidance of experienced Data Scientists.
- Perform exploratory data analysis and data risk assessments to identify data gaps, quality issues, potential bias, confounding factors, outliers, and limitations that may affect algorithm development or evaluation.
- Define and document data needs for training, tuning, and evaluation, including relevant subgroups and acceptance criteria, so that data cohorts are representative of the intended use and aligned with the product and regulatory strategy.
- Support the specification of annotations, acceptance criteria, and quality controls; collaborate with qualified internal or external suppliers on sourcing data and annotations.
- Curate, version, maintain, and document representative data cohorts and ensure data provenance, traceability, quality, and fitness for purpose together with data management stakeholders.
- Design, implement, train, tune, and technically characterize AI/ML algorithm components using appropriate machine learning, deep learning, statistical, or generative AI methods.
- Define scientifically justified metrics, evaluation methods, test cases, and acceptance criteria to assess stand-alone algorithm performance, including relevant aspects such as accuracy, robustness, reliability, generalizability, explainability, bias, and subgroup performance.
- Implement and execute reproducible bench tests on representative evaluation cohorts and relevant subgroups as part of the verification and validation strategy.
- Analyze and interpret evaluation results, investigate failure modes, document residual limitations, and derive evidence-based recommendations for algorithm improvement and product decisions.
- Document development and evaluation activities and contribute to Data Science work products such as data management information, annotation specifications, scientific evaluation test cases, and scientific evaluation reports.
- Provide scientific evidence and technical input for product reviews and regulatory submissions and support responses to questions concerning data, annotations, evaluation methods, and stand-alone algorithm performance.
- Support the design and execution of clinical investigations, reader studies, or other studies when additional evidence is required by the product or regulatory strategy.
- Collaborate with software engineering and MLOps functions to establish reproducible pipelines, versioned artifacts, automated tests, controlled deployment, and integration of algorithm components into the product.
- Help define and implement monitoring metrics for detecting data anomalies, outliers, and algorithm performance drift across the product lifecycle.
- Regularly review advances in Data Science and AI and assess their applicability to the development and evaluation of AI/ML products.

Qualification:

Master's degree in Computer Science, Data Science, Statistics, Mathematics, Biomedical Engineering, Medical Informatics, or a related quantitative or engineering field. A Ph.D. is welcome but not required. The candidate should have completed university-level coursework or equivalent practical training in:

- Probability, statistics, and experimental design
- Artificial intelligence, machine learning, and deep learning
- Linear algebra, optimization, and numerical methods
- Data structures, algorithms, and software programming
- Model evaluation, scientific methodology, and interpretation of empirical results

Experience level:

This is a junior R&D position. Relevant experience may have been gained through academic research, a thesis, internships, working-student assignments, or up to approximately three years of professional work in Data Science, machine learning, AI research, or AI/ML product development. Demonstrated hands-on experience in developing and scientifically evaluating AI/ML models is more important than a specific number of years.

Desired Knowledge & Experience:

- Sound understanding of machine learning, deep learning, statistical analysis, experimental design, and model evaluation.
- Active programming skills in Python and practical experience with libraries and frameworks such as NumPy, Pandas, SciPy, scikit-learn, PyTorch, TensorFlow, or Keras.
- Hands-on experience with the development workflow of AI/ML models, including data preparation, training, tuning, evaluation, error analysis, reproducibility, and documentation.
- Ability to define meaningful performance metrics, evaluation cohorts, subgroup analyses, test methods, and acceptance criteria for the intended use of an AI/ML algorithm.
- Understanding of data quality, representativeness, bias, annotation quality, data provenance, and traceability in AI/ML development.
- Knowledge of Machine Learning Operations (MLOps) concepts and tooling, including source and data versioning, reproducible pipelines, experiment tracking, automated testing, CI/CD, deployment, and monitoring.
- Hands-on exposure to current Large Language Models and Generative AI, including foundation models, prompting, retrieval-augmented generation, and model adaptation or fine-tuning.
- Knowledge of Agentic AI, including tool-using models, multi-step workflows, orchestration, memory, and autonomous or semi-autonomous decision paths, is an advantage.
- Experience or strong interest in the systematic evaluation of LLM, Generative AI, and Agentic AI systems, including task-specific benchmarks, groundedness and factuality, robustness, reproducibility, human evaluation, safety and guardrail testing, tool-use correctness, and analysis of failure modes.
- Experience with databases and query languages such as SQL; knowledge of Kusto Query Language, Snowflake, or graph databases such as Neo4j is an advantage.
- Exposure to cloud-based data and AI environments, preferably Microsoft Azure, such as Azure Blob Storage, Azure Databricks, Azure Data Factory, Azure Machine Learning, Azure Monitor, or Azure DevOps.
- Experience with structured and unstructured data, for example machine logs, event data, Parquet files, text data, or multimodal data.
- Knowledge of healthcare, medical technology, clinical data, or interoperability standards such as DICOM is preferred.
- Basic project management skills and familiarity with working in cross-functional R&D or product-development teams.

Required Soft Skills & Other Capabilities:

- Scientific mindset, strong analytical ability, and attention to detail.
- Commitment to objective evidence, reproducibility, traceability, data integrity, and responsible AI development.
- Curiosity and resilience to investigate unexpected results, challenge assumptions, and improve models based on evidence.
- Collaborative approach to working with Data Scientists, domain experts, software engineers, product management, and quality and regulatory stakeholders.
- Ability to work independently within an agreed scope and to seek guidance when technical, scientific, product, or regulatory implications require alignment.
- Clear written and verbal communication skills in English, including the ability to explain methods, results, uncertainties, limitations, and risks to technical and non-technical stakeholders.