Kerja Sepenuh Masa, Data Scientist di Oncoshot - Maukerja

## 1. Role Summary

Oncoshot turns unstructured clinical documents (clinical notes, pathology, radiology and genomic

reports) into a structured, research-grade record of each institution’s cancer population:

diagnoses, histology, TNM and overall stage, biomarkers, treatments and disease status. Hospitals

and research partners rely on that record only as far as its quality can be shown. The Data

Scientist (Evaluation & Data Quality) owns the science of measuring it: how gold-standard data is

sampled and annotated, how accuracy is estimated and compared between releases, how far model

confidence can be trusted, and how complete and consistent the record is at each institution.

This is a hands-on, statistics-driven role. The engineers build and ship the pipelines; the Data

Scientist defines how quality is measured, what counts as a real improvement and where the bar

sits. Patient-derived data never leaves the institution, so analyses run inside each hospital’s

environment or on de-identified and aggregated results. To see how we measure extraction quality,

look at our open-source evaluation framework, llmvalidate

(*************), which this role works with and extends.

## 2. Key Responsibilities

- **Design the gold standard:** Sample evaluation data stratified by document type, institution and

  rare but clinically important values, size each sample for the precision required, and decide

  when it must be refreshed.

- **Run annotation as a measurement process:** With the clinical validation team, write annotation

  guidelines, measure inter-annotator agreement, adjudicate, and turn recurring disagreements into

  sharper field definitions.

- **Own the evaluation methodology:** Specify how each field is scored, including multi-valued and

  hierarchical fields such as staging and biomarkers; report precision, recall, F1 and F2 with

  confidence intervals; and define the statistical tests that let release gates separate a real

  change from noise.

- **Calibrate confidence:** Measure how well confidence scores predict correctness, and set the

  thresholds that send results to human review, balancing accuracy against review workload.

- **Measure the quality of the record:** Monitor completeness, internal consistency (e.g. TNM against

  overall stage) and conflicts between sources at each institution, detect drift when documents,

  templates or case mix change, and report which fields are fit for which purpose.

- **Protect patient data:** Design analyses that run where the data lives, use synthetic data for

  development, keep identifiers out of reports, logs and tickets, and contribute to ISO 27001

  evidence.

## 3. Preferred Qualifications

We don’t expect any one candidate to have all of these; strong Python, statistical rigour and

curiosity about clinical data matter most.

- Bachelor’s or Master’s degree in Statistics, Biostatistics, Data Science, Epidemiology,

  Biomedical Informatics, Computer Science, or a related quantitative field.

- Experience as a data scientist or applied statistician on real-world data, with strong hands-on

  Python (e.g. pandas, SciPy, statsmodels, scikit-learn) and SQL.

- Measurement and inference: sampling design and sample-size estimation, confidence intervals and

  the bootstrap, hypothesis testing and multiple comparisons, and inter-annotator agreement (e.g.

  Cohen’s or Fleiss’ kappa).

- Evaluating machine-learning, NLP or LLM outputs against human-labelled data, including error

  analysis, calibration and the design of annotation studies.

- Reproducible practice: Git, code review, tested analysis code and versioned datasets; cloud

  experience (AWS, Azure or GCP) is an advantage.

- Experience with clinical or health data is strongly preferred (e.g. EMR/EHR, cancer registries,

  ICD-O-3, AJCC TNM, SNOMED CT, real-world data-quality frameworks), as is familiarity with

  health-data privacy frameworks (PDPA, GDPR, HIPAA).

## 4. Core Competencies

- **Pragmatic measurement:** Annotation is expensive; spends it where it changes a decision and

  designs samples that answer the question with as little labelling as possible.

- **Clinical precision:** Turns ambiguous medical concepts into exact, measurable definitions, and

  asks a clinician when the answer is not clear-cut.

- **Communication across audiences:** Explains quality, uncertainty and trade-offs credibly to

  clinical, commercial and engineering stakeholders.

## 5. Success Measures (First 3 Months during Probation Period)

- Published an assessment of the existing gold-standard datasets and evaluation methodology —

  coverage, sample size and annotation agreement — with a prioritised plan agreed with the CTO.

- Introduced confidence intervals and a statistical release criterion into the regression gate of

  at least one extraction pipeline.

- Delivered a first completeness and consistency report for the structured record at one or more

  institutions, with thresholds agreed with the CTO.

## 6. Working Model

- **Probation period:** 3 months, followed by conversion to permanent staff.

- **Reporting line:** Reports to the Chief Technology Officer.

- **Location & arrangement:** Hybrid, based in Singapore (one-north), with occasional on-site work at

  partner hospitals in Singapore, including on data that must stay within the institution.

- **Eligibility:** Open to Singapore Citizens and Permanent Residents only.

- **Team structure:** Works with the AI and data pipeline engineers who own the extraction and

  evaluation code, with clinicians and the clinical validation team on what “correct” means for

  each field, and with product teams on the quality customers need.

- **Cadence:** Agile/Scrum, with sprint planning, refinement and reviews.

## 7. Salary Range

• Indicative monthly gross salary: SGD6,000 – SGD10,000, based on experience and seniority.