AI Product Lead

Role Purpose

Lead the product vision, delivery, and adoption of an AI-enabled knowledge base for internal banking operations—turning policies, procedures, risks, controls, and processes into trusted, versioned knowledge assets (including knowledge graphs). You will sit at the intersection of product, engineering, and domain stakeholders, shaping the roadmap for Enterprise Knowledge and Agent Systems to transform structured and unstructured internal content into actionable intelligence that powers internal-facing agentic systems and automation.

Key Responsibilities

Product Leadership & Delivery

Own the Product Strategy: Lead the product strategy and roadmap for the Enterprise knowledge platform (extraction, representation, validation, and consumption).
Translate Internal Needs: Convert business requirements (e.g., policy interpretation, control mapping, risk insights) into clear product requirements, user stories, and acceptance criteria.
Prioritise Workstreams: Drive delivery across multiple streams—extraction, ontology/graph, quality, and platform integration—ensuring measurable outcomes for internal efficiency.
Define Success: Track metrics including coverage, accuracy, freshness, latency, and internal adoption rates.

Knowledge Extraction & Structuring (LLM-enabled)

Internal Data Pipelines: Guide the development of pipelines that convert unstructured internal content into structured artefacts using LLMs and NLP techniques.
Standards & Evaluation: Set standards for prompt engineering, evaluation, and regression testing of extracted internal artefacts.
Entity Resolution: Oversee approaches to ensure consistent identifiers across siloed internal systems (e.g., Core Banking, ERP, and procedural documentation).

Knowledge Representation & Graph/Ontology Governance

Representation Design: Lead design decisions regarding ontology alignment, graph modelling patterns, and versioning strategies.
Adherence to Standards: Ensure all knowledge models align with broader ontology standards.
Trust & Lineage: Establish versioning, lineage, and change management so internal consumers can audit and trust system outputs.

Quality, Risk, and Controls

Quality Gates: Implement automated and manual quality checks for extraction and graph population.
Risk Partnership: Partner with risk, compliance, and governance stakeholders to ensure the platform meets control expectations for internal commercial banking content.
Responsible AI: Drive practices that ensure transparency, traceability, and safe deployment patterns for internal agents.

Platform Integration & Enablement

Agentic Workflows: Ensure the knowledge base is consumable via APIs to support internal use-cases such as search, Q&A, and agent-led orchestration.
Internal Application Support: Collaborate with teams integrating LLM agents into internal-facing products, ensuring high performance and reliability.
Onboarding & Adoption: Create enablement materials (playbooks, documentation) to accelerate the adoption of agentic systems across internal banking teams.

What You'll Bring (Required)

Strong Python engineering background with solid software design and engineering practices.
Proven experience integrating LLMs into production software (including evaluation, monitoring, and iteration).
Hands-on experience with:
Methods for converting unstructured to structured data using LLMs.
Prompt engineering and prompt lifecycle management.
API integration and service-oriented design.
Git-based development workflows.
Working knowledge of NLP concepts and practical application in information extraction.
Knowledge graph & Ontology modelling: Experience with knowledge graph design & development and ontology-driven modelling (enough to lead decisions and guide specialists).
Cross-functional leadership: Ability to lead delivery by balancing internal stakeholder needs, technical constraints, and delivery pace.
Nice to Have

Knowledge of Commercial Banking policies, procedures, risks, controls, and process documentation.
Deeper expertise in ontology engineering or knowledge graph platforms (e.g., Neo4j, RDF/OWL).
Deeper expertise in NLP pipelines and information extraction evaluation methodologies.

Ways of Working/Behaviours

Comfortable with Ambiguity: Able to turn broad enterprise goals ( turn “we need better policy answers” ) into a deliverable product plan.
Collaborative Leadership: An inclusive style that aligns engineers, internal SMEs, and governance partners.
Strong Ownership Mindset: Proactively setting direction and unblocking delivery to ensure the platform is used and trusted.

Example Deliverables (First 90 Days)

A prioritised roadmap covering extraction, ontology/graph, quality gates, and consumer APIs.
A measurable quality framework tailored to internal banking documentation (precision/recall, entity resolution KPIs).
A reference architecture for LLM-assisted extraction + knowledge graph population.
A pilot enterprise use-case (e.g., internal policy co-pilot) integrated end-to-end.