Senior Data Scientist AI-Driven Healthcare Entity Linking

Sii Poland; Wrocław; On-site; PLN 260,000 - 360,000; Full time

Job summary

Sii Poland is seeking a Senior Data Scientist to join an AI-driven operating system for healthcare marketing. You will design and implement entity extraction, linking, and resolution at scale, reconciling large and fragmented datasets into reliable entities and attributes.

You will own the linking pipeline end-to-end, apply probabilistic scoring, and validate quality using precision, recall, calibration, and error analysis while aligning with controlled medical vocabularies.

Qualifications

• 5+ years of production ML or data-science experience.
• Hands-on entity linking and resolution at scale.
• Production-grade Python with typing in code.
• Strong SQL, FastAPI, Docker, pytest, Git, CI/CD knowledge.
• Experience with probabilistic methods for match scoring and calibration.
• Graph ML and embeddings experience.

Responsibilities

• Extract entities and attributes from heterogeneous sources.
• Design probabilistic entity linking and deduplication across large healthcare datasets.
• Own the linking pipeline end to end with graph-based scoring.
• Define link quality metrics: precision, recall, calibration, error analysis.
• Link entities to controlled medical vocabularies for consistency.
• Apply graph representations and ML techniques for linking and clustering.
• Maintain data and graph estates including ETL transformations.
• Improve algorithms as data sources evolve.

Skills

Entity linking

Graph ML

ML engineering

Python

SQL

Data quality

Embeddings

GNNs

Tools

FastAPI

Docker

pytest

Git

CI/CD

Snowflake

BigQuery

Databricks

Neo4j

Amazon Neptune

Job description

Sii Poland is seeking a Senior Data Scientist to join an AI-driven operating system for healthcare marketing. You will design and implement entity extraction, linking, and resolution at scale, reconciling large and fragmented datasets into reliable entities and attributes.

You will own the linking pipeline end-to-end, apply probabilistic scoring, and validate quality using precision, recall, calibration, and error analysis while aligning with controlled medical vocabularies.