Senior AI Platform Architect

Sumitomo Pharma•Marlborough, MA

2d•Remote

About The Position

Sumitomo Pharma Co., Ltd., is a global pharmaceutical company based in Japan with operations in the U.S. (Sumitomo Pharma America, Inc.), focused on addressing patient needs in oncology, urology, women's health, rare diseases, cell & gene therapies and CNS. With several marketed products and a diverse pipeline of early- to late-stage investigational assets, we aim to accelerate discovery, research, and development to bring novel therapies to patients sooner. The Senior AI Platform Architect is a senior individual contributor responsible for partnering with business and technology stakeholders to identify, define, and prioritize AI opportunities, then architect and deliver working implementations in a secure, reliable, and compliant manner. This role designs and evolves the enterprise AI platform and delivery patterns spanning traditional ML and modern LLM-based solutions, enabling teams to move from use case to production through reusable components, MLOps/LLMOps practices, and automation.

Requirements

BS in Information Technology or related field.
8+ years of experience designing and delivering cloud platforms and/or production software, including 3+ years building or operating AI/ML solutions (traditional ML and/or LLM-based systems) in a regulated environment.
Experience working with multi-tiered ticket handling/resolution systems.
Strongly motivated, well-organized, and shows professional initiative.
Takes ownership of assignments, can work both independently and as part of the team.
Ability to work in a fast-paced and challenging environment.
Excellent organizational skills with the ability to manage multiple tasks simultaneously.
Exceptional written and oral communication skills.
Ability to work responsibly with minimal supervision and with a sense of urgency while maintaining the proper judgment on when to inform management on an issue and escalate concerns or customer escalations.
Display ability to solve problems.
Deep knowledge of AI/ML fundamentals and LLM architectures and patterns, with strong hands-on experience in Python and Terraform; familiarity with CI/CD and major cloud platforms (AWS and/or Azure) for production AI workloads.

Nice To Haves

Customer service mindset.
Attentive problem solver.

Responsibilities

Lead end-to-end AI use case delivery: facilitate discovery with business partners, define problem statements, success metrics, data needs, constraints, and deliver working prototypes that mature into production implementations.
Architect and evolve an enterprise AI/ML/LLM platform including reference architectures, reusable patterns, and paved roads for model training, evaluation, deployment, and monitoring (MLOps/LLMOps).
Design and implement LLM solutions such as RAG, tool/function calling, prompt management, agent workflows, and safety guardrails; select appropriate models and approaches based on use case requirements.
Define data and integration patterns for AI solutions, including feature access, document ingestion pipelines, vector search, APIs/eventing, and secure connectivity to internal systems.
Build automation and infrastructure-as-code to provision and operate AI workloads using Terraform (e.g., cloud resources, IAM, networking, storage, observability) and Python (services, pipelines, tooling).
Establish and enforce engineering standards and best practices for AI systems: reproducibility, testing, model evaluation, documentation, cost controls, and operational readiness.
Partner with security, privacy, and compliance teams to implement governance controls for AI (e.g., access management, data handling, auditability, model risk management, validation requirements).
Own reliability and performance for AI platform services by designing for resilience and scalability, instrumenting telemetry, and driving incident response and root cause remediation.
Enable and guide application and data teams through design reviews, hands-on pairing, internal documentation, and reference implementations.
Communicate architecture decisions and tradeoffs clearly to technical and non-technical stakeholders; manage competing priorities to deliver outcomes aligned with business value.