Data Scientist

At Pretius we are looking for an experienced Data Scientist to join a project for a client in the insurance industry. The project combines classic machine learning with Generative AI agents, built entirely on Oracle Cloud Infrastructure (OCI). You will help build solutions that support underwriting, actuarial analysis and fraud, waste and abuse (FWA) detection, with a strong focus on security, traceability and client data isolation.

Project / Role
• Build ML and statistical models (e.g. risk scoring, fraud detection) for insurance clients on OCI.
• Prepare data and manage the model lifecycle: training, batch scoring, deployment and drift monitoring.
• Run daily batch scoring and large-scale data processing.
• Design GenAI agents with RAG over insurance documentation and governed access to data through SQL.
• Build and evaluate the vector search layer behind the agents.
• Implement guardrails and traceability for GenAI solutions.
• Ensure security and isolation of client data in a multi-tenant setup.

Requirements
• 3+ years in statistics and machine learning.
• Strong PL/SQL skills and hands-on work with Autonomous Database (ADW) on OCI.
• Practical knowledge of OCI Data Science: notebook sessions, Jobs, ML Pipelines, conda environments, Model Catalog (versioning, metadata) and Model Deployment.
• Model monitoring for drift and performance.
• Working with Autonomous Database and Object Storage from Data Science, with a clear view on when to train in-database vs. externally.
• Scheduling and orchestration of batch scoring on daily-refreshed data.
• Background in LLM- and RAG-based solutions, ideally on OCI Generative AI, including hosting or calling LLMs in governed agents.
• Ability to build agents that query governed data through SQL and function-calling tools instead of answering numbers from model memory.
• Solid grasp of retrieval fundamentals: chunking, embeddings, retrieval quality evaluation.
• Solid grasp of guardrails and traceability: prompt-injection controls, citations, trace logging of every tool call and answer, human approval for consequential actions.
• Good command of cloud security: IAM policies, Vault, private networking, Logging and Monitoring.
• Fluent English (min. B2).

Nice to have
• Oracle Machine Learning (OML) and Select AI, including where Select AI fits and where it doesn’t.
• Familiarity with OCI Generative AI Agents: agents, endpoints, sessions and knowledge bases over documents such as policy wordings, underwriting guidelines, actuarial memos and FWA investigation playbooks.
• Exposure to Oracle AI Vector Search or OpenSearch.
• Exposure to OCI Data Flow (Spark).
• Insurance domain knowledge (underwriting, actuarial, FWA).
• Background in multi-tenant solutions that keep client data and models isolated.
 

What do we offer?
• We focus on long-term relationships based on fair principles and reliability.
• Co-financing of the Multisport card and Medicover private healthcare.
• Modern office available.
• Team bonding activities, internal courses, conferences, certifications.