AI Engineer - Lead

Job description

**Job Title:** TWIN AI Engineer **Department** : Production STA – AIOPS **About Business line/Function** **:** The BNP Paribas Group's IT department (IT GROUP) aims to ensure the overall consistency of the Information System (IS) with the Group's strategic objectives, while improving the management and control of operational risks. The Role Of IT Includes, Among Other Things, Providing Enhanced Customer Experience, Accelerating Digitalization, And Optimizing Operational Efficiency. Thus, IT GROUP Has Modernized The Information Systems By Deploying Digital Levers To Support The Digitalization Of The Group's Businesses And Functions: - The delivery of digital levers (Cloud, API, Data, Digital Working, Tech watch) is effective, and the IT Market Place enables their deployment on a Group-wide scale, - The security of the IT environment is evolving with a focus on cybersecurity risks and IT resilience to meet increasingly stringent regulatory requirements, - The robustness of the Information Systems is improving with the implementation of standardized operational methods, - Human capital remains at the center of our priorities, - Operational efficiency is growing with a focus on process automation and industrialization, - All these advances contribute to the Group's ambition to reduce its environmental footprint. Thanks to these solid foundations, a new phase of the IT strategy is opening up, enabling the Group to address a dual challenge: accelerating the adoption of innovative solutions around data and Artificial Intelligence (AI), and ensuring the interoperability and security of Information Systems. To Meet These Challenges, IT GROUP Is Adapting Its Organization And Remaining In Line With That Of The Bank. This Evolution Is Based On Five Main Levers: - Technological innovation through data and AI: ensuring excellence in innovation through research and development of skills, relying on specific governance. This involves structuring technical foundations to the state of the art in order to accelerate experimentation, the construction of innovative solutions, and their adoption on a large scale by the Business and Functions, - Cybersecurity: continuing to protect the Bank and its customers in terms of Cybersecurity and Cyber Fraud, in line with regulatory requirements, - Infrastructure and Production: creating synergies, accelerating industrialization, emphasizing interoperability, and improving the resilience of Information Systems, while controlling risks, - Payment and Card Services: meeting the challenges of the industry by ensuring optimal alignment and end-to-end execution of strategic programs among all Group stakeholders, - Operational excellence: gaining efficiency, particularly in financial terms, to continue optimizing expenses. ****Position Purpose:**** **AIOPS** The Twin Knowledge Engineering Specialist will build and maintain the knowledge pipelines that allow Twin agents to produce reliable, explainable and grounded answers. **The Role Will Contribute To:** - ingestion of operational knowledge; - metadata enrichment; - semantic structuring; - RAG and GraphRAG preparation; - knowledge graph population; - knowledge quality controls; - source traceability; - certification of exposed sources; - detection of obsolete, duplicated or contradictory knowledge; - continuous improvement of knowledge retrieval. **The Key Principle Is:** **No reliable Twin without reliable knowledge engineering.** ****Responsibilities**** Direct Responsibilities - Engineer knowledge ingestion pipelines Build and operate the technical pipelines used to ingest knowledge into Twin. **Responsibilities Include:** - Implement ingestion workflows for documents, procedures, FAQs, incident knowledge, change knowledge, infrastructure knowledge and operational standards. - Extract content from structured and unstructured sources. - Normalize formats before ingestion. - Apply metadata enrichment. - Manage document versioning and lifecycle metadata. - Prepare content for RAG indexing. - Prepare content for graph-based representation when relevant. - Detect ingestion errors, incomplete documents and unsupported formats. - Implement automated quality checks before exposure to Twin agents. - Structure knowledge for RAG and GraphRAG Transform raw knowledge into AI-usable knowledge. **Responsibilities Include:** - Define and apply chunking strategies. - Implement semantic tagging. - Add metadata such as owner, source system, validity date, confidentiality level, domain, franchise, CI, application, service and operational scope. - Optimize retrieval quality through indexing strategy. - Support hybrid search, semantic search and graph-assisted retrieval. - Link knowledge chunks to entities in the knowledge graph. - Maintain relationships between applications, infrastructure components, procedures, incidents, changes and services. - Improve grounding and source citation capabilities. - Build and maintain knowledge graph content Support the creation and maintenance of graph-based knowledge used by Twin. **Responsibilities Include:** - Model key IT production entities and relationships. - Populate graph structures from documentation, CMDB, inventories and operational sources. - Support entity resolution and deduplication. - Detect inconsistent relationships. - Contribute to graph quality checks. - Support GraphRAG use cases. - Validate graph traversal results used by agents. - Work with Twin Core and graph specialists on ontology, schema and graph evolution. - Implement knowledge quality controls Develop controls to ensure that Twin consumes trusted and usable knowledge. **Responsibilities Include:** - Implement automated checks for missing metadata. - Detect obsolete documents. - Detect duplicated or conflicting content. - Flag documents without owner or validation date. - Track source freshness. - Monitor knowledge coverage by domain or fran