Data Engineer

Key responsibilities

• Design and build ingestion pipelines for structured data, unstructured content (documents, PDFs), and metadata.
• Build reusable connectors to enterprise systems, catalogs, content repositories, and third-party or licensed sources via APIs.
• Model and build raw-to-mart data layers that serve analytics and AI use cases.
• Implement metadata extraction, enrichment, and search indexing, including semantic and vector indexes for RAG.
• Set up and manage vector databases and knowledge repositories used by LLM agents.
• Implement data quality rules, profiling, scoring outputs, and exception handling.
• Register lineage and maintain source registries, version tracking, and refresh controls.
• Design data stores for signals, findings, audit trails, and user feedback loops.
• Apply security and governance controls: RBAC, PII/sensitivity flagging, and approved data handling.
• Support SIT/UAT data validation, defect fixes, and production release activities.

Required skills

• Strong Python and SQL; solid data modeling (dimensional and normalized).
• Hands-on experience with a modern data platform such as Databricks, Snowflake, or Azure/AWS data services.
• Pipeline orchestration and transformation (Spark, Airflow, ADF, dbt, or similar).
• API-based integration (REST), JSON handling, and incremental/CDC ingestion patterns.
• Data quality frameworks and testing practices for pipelines.
• Version control (Git) and CI/CD for data workloads.

Preferred skills

• RAG data preparation: chunking, embeddings, vector databases (Azure AI Search, pgvector, Pinecone, or similar).
• Unstructured content processing: text extraction, OCR, document parsing.
• Metadata management, data catalogs, ontologies, or knowledge graphs (for example Neptune or other graph databases).
• Experience supporting LLM or agentic applications with grounded, traceable data.
• Life sciences data exposure (commercial, medical, regulatory, or launch data) and regulated-data handling.
• Cloud certification (Azure Data Engineer, Databricks, AWS, or Snowflake).