Director, AI Systems Foundations & Reusable Capabilities

Salary Range:

- Job Description Summary Director, AI Systems Foundations & Reusable Capabilities#LI-Hybrid

Location:

London Novartis is unable to offer relocation support for this role: please only apply if this location is accessible for you.

The Director, AI Systems Foundations & Reusable Capabilities is a senior technical AI leadership role within Data Science & AI, responsible for establishing the reusable foundations that enable Development AI systems to be built faster, more consistently, and with greater technical leverage across Novartis.

The role identifies common technical patterns across AI initiatives and shapes the long-term evolution of Development's AI capability stack across agentic AI, retrieval and knowledge systems, simulation and digital twins, MLOps, LLMOps, Agent Ops, and AI lifecycle management. It translates proven architectures, components, engineering practices, and design decisions into reusable AI capabilities that reduce duplication and ensure learning from one system strengthens the next.

This role owns reusable AI system foundations and technical AI capability assets, not product roadmaps, business adoption, portfolio governance, infrastructure operations, or enterprise architecture. It partners closely with DSAI domain leaders, the AI Systems Reliability Director, Product teams, Strategy & Governance, and DDIT to ensure AI capabilities are reusable, interoperable, technically coherent, and continuously improved.

Job Description Major Accountabilities AI Foundations & Reference Architectures Define reusable architectures, design patterns, and reference implementations for AI systems across Development.

Establish common approaches for agentic AI, retrieval-augmented generation, knowledge systems, simulation, digital twins, and human-in-the-loop AI.Provide technical guidance on reuse, extension, and evolution of AI capabilities.

Ensure AI systems are designed for modularity, interoperability, scalability, and maintainability.

Reusable AI Capabilities & Engineering Practices Build and maintain reusable AI components, frameworks, templates, and capability assets.

Define AI engineering practices and reference approaches for MLOps, LLMOps, and Agent Ops.

Establish standards for lifecycle management, versioning, orchestration, observability, and technical documentation.

Identify recurring technical needs and convert them into reusable organizational capabilities.

Capability Scaling & Technical Enablement Partner with domain leaders to identify successful patterns that can be reused across Development.

Translate project-level solutions into reusable architectures, components, and implementation blueprints.

Guide AI engineers and data scientists in designing systems that can be reused and extended across domains.

Promote technical consistency while preserving flexibility for innovation.

Capability Learning & Evolution Capture architectural, engineering, and implementation learnings from AI initiatives and transform them into reusable assets.

Build mechanisms for sharing AI patterns, reference designs, and technical best practices across DSAI.Partner with the Reliability Director to connect reusable capabilities with evaluation evidence, reliability insights, and production readiness criteria.

Continuously evolve the Development AI capability base as technologies and practices mature.

Key Performance Indicators
· Percentage of strategic AI systems leveraging reusable AI patterns or reference architectures.
· Number of reusable AI capabilities adopted across multiple domains.
· Reduction in duplicate technical solutions across the portfolio.
· Reduction in time required to build new AI systems through capability reuse.
· Number of project learnings converted into reusable capability assets.
· Technical satisfaction and adoption of reusable capabilities within DSAI.Minimum Requirement:
Work Experience
· 10+ years leading AI, machine learning, AI engineering, or technical architecture initiatives.
· Demonstrated experience designing reusable AI architectures, platforms, frameworks, or technical capabilities.
· Deep expertise in modern AI systems, including LLMs, RAG, agentic systems, orchestration,…