VP, Chief AI Architect

Role Summary

Pfizer is seeking a Vice President, Chief AI Architect to define and steward the enterprise AI architecture vision, roadmap, and governance that enable breakthroughs at scale across R&D, Manufacturing, and Commercial. This role brings integrated, end-to-end thinking across data, models, platforms, and products; curates innovation from a strong external network; and ensures secure, reliable, cost-effective patterns for AI solutions (including LLMs and agentic systems) in a regulated environment. The Chief AI Architect partners closely with the Head of AI CoE to ensure that reference architectures, standards, and guardrails are translated into scalable, reusable capabilities. Together, they accelerate adoption, improve reliability and time-to-value, and uphold Responsible AI principles.

Responsibilities

• Define the target‑state AI architecture (data, model, application, and infrastructure layers) that integrates advanced analytics, ML/LLM, knowledge/semantic technologies, and operational systems.
• Establish the North Star for foundational capabilities: RAG and retrieval pipelines, agents/orchestration, vector search, feature stores, model registries, observability, evaluation, safety layers, etc.
• Set architecture principles that balance innovation speed with compliance, reliability, and total cost of ownership.
• Publish reference architectures and blueprints for priority use cases (e.g., scientific discovery assistants, GxP‑impacted automation, manufacturing QA, field engagement copilots).
• Define LLMOps/MLOps standards (model lifecycle, evaluation, red‑teaming, monitoring, rollback, drift, lineage, documentation).
• Codify security, privacy, and Responsible AI guardrails: data minimization, isolation patterns, PII/PHI handling, human‑in‑the‑loop, explainability, auditability, model risk controls.
• Own the enterprise AI architecture roadmap; align with business strategy and portfolio funding.
• Chair an AI Architecture Review Board (AARB) and design authorities that provide fast, pragmatic guidance and approvals.
• Manage technology lifecycle (emerging → adopt → scale → retire) for AI frameworks, model classes, toolchains, and platforms.
• Maintain a strong external network (hyperscalers, model labs, hardware vendors, startups, academia, standards bodies) to scout, evaluate, and curate innovations.
• Run evidence‑based proofs‑of‑value and bake successful patterns into the reference stack; shape build/partner/buy decisions with the Head of AI CoE and Procurement.
• Represent Pfizer’s interests in industry consortia and standards discussions; encourage selective open‑source contribution where it benefits the enterprise.
• You set the blueprint; the CoE builds/operates. Co‑own the platform backlog prioritization and ensure reference patterns → productized capabilities.
• Define SLAs/SLOs, performance benchmarks, and cost guardrails in collaboration with the CoE and SRE/FinOps.
• Jointly drive developer enablement: SDKs, templates, golden paths, sandboxes, and documentation.
• Embed model risk management, validation evidence, and audit‑ready documentation into patterns—fit for GxP, 21 CFR Part 11, GDPR/HIPAA contexts as applicable.
• Institutionalize AI safety: pre‑production evaluations, content safety, adversarial testing, policy enforcement, incident response playbooks.
• Promote API‑first and event‑driven integration between AI services and enterprise systems; enable semantic/knowledge layers to unify context across domains.
• Maximize reuse via shared components (prompt libraries, evaluation suites, connectors, datasets, ontologies), tracked through measurable reuse rates.
• Set performance engineering practices for training, fine‑tuning, and inference (e.g., quantization, distillation, caching, batching).
• Partner with Infra/Cloud/HPC on capacity planning (GPU/accelerator utilization), autoscaling, and cost/per‑inference optimization.
• Build an AI Architecture Guild that mentors domain architects and product teams.
• Develop playbooks, training, and office hours to raise architectural quality and speed across the enterprise.

Qualifications

• Required: BS/BA degree required, higher degree preferred or relevant experience, 15+ years in architecture or advanced engineering leadership, with 7+ years designing AI/ML platforms and solutions at enterprise scale.
• Required: Demonstrated mastery across LLMs/foundation models, retrieval/RAG, agents/orchestration, evaluation, model safety, and LLMOps/MLOps.
• Required: Deep experience in regulated environments (life sciences/healthcare or equivalent), including validation, auditability, and documentation rigor.
• Required: Proven ability to create reference architectures and standards and drive adoption through governance that enables speed (not bureaucracy).
• Required: Strong external network and a track record of curating innovation (ecosystem scouting, PoVs, build/partner/buy).
• Required: Hands‑on credibility with modern stacks: vector databases, feature stores, model registries, observability, event‑driven and API‑first integration, cloud/HPC, and performance engineering for training and inference.
• Required: Exceptional influence and storytelling skills; able to align senior stakeholders and simplify complex trade‑offs.
• Preferred: Prior leadership of enterprise or domain architecture for AI‑heavy portfolios.
• Preferred: Contributions to open‑source, standards, reference implementations, or published thought leadership.
• Preferred: Familiarity with data mesh/semantic layers/knowledge graphs, and FinOps/SRE practices for AI platforms.

Education

• BS/BA degree required; higher degree preferred.

Additional Requirements

• Travel up to 20% may be required for business activities.