The company's payments business serves customers in more than 200 countries and territories, with a focus on secure, simple and accessible digital payments.
The company is hiring a Lead AI Engineer for a small AI technology team led by one of its Distinguished Engineers. The team's first work covers agentic technologies, including the Model Context Protocol (MCP) at scale, and the use of open weight AI models. The team works with key technology stakeholders across the organization.
This is a hands-on engineering role. You will build the platform capabilities that let the company adopt MCP and open weight models safely and at scale, working alongside more senior engineers who set the wider technical direction.
You will help build the platform patterns that let teams expose tools and data to agents over MCP, and that let the enterprise serve, evaluate and govern open weight models. You will help shape key technical decisions on gateway and connectivity, identity and access controls, model-serving infrastructure and evaluation, then turn them into reliable, production-grade services. You will work across the MCP estate and the open weight model working group, with the team's senior engineers and with central technology, security and governance stakeholders.
Responsibilities
Help design and build the parts of the company's enterprise MCP deployment, such as the security gateway, the MCP registry that records what exists and who owns it, and reference server implementations teams can build from. Take them from design to running production services.
Implement identity, authentication and authorization patterns for MCP across the estate, working with senior engineers and central security and identity teams. Agents, tools and data must be accessed securely and be auditable end to end.
Help define and apply the standard for internal and external MCP servers, covering how tools and data are exposed, discovered and versioned, and which inspection, data-protection, rate-limiting and kill-switch controls the gateway enforces.
Build and run open weight model serving. Stand up and scale self-hosted inference, package models for deployment, and help make it possible to move workloads between models and providers for cost, flexibility and resilience.
Build evaluation and testing tooling that benchmarks candidate open weight models on quality, cost, latency and safety, and produces reproducible evidence for governance and security reviews.
Apply MLOps practices that keep the platform reliable: CI/CD, infrastructure-as-code, observability, and model and service monitoring, so the platform stays maintainable in production.
Work with central security, privacy, legal and AI governance teams, the MCP SIG and the open weight working group, and implement their requirements correctly in the platform's architecture.
Track fast-moving developments such as new MCP features, frontier open weight releases and related agentic protocols, and help the team decide what to adopt, prototype or wait on.
All About You
Skills, experience and attributes that would make you a strong candidate:
Strong knowledge of modern AI: LLMs and foundation models, agents, and the tools and protocols, such as MCP, that connect them to enterprise systems.
Good software engineering skills. You write well-tested, production-quality code and build services other teams can depend on, not one-off scripts or prototypes.
Solid MLOps experience: CI/CD, infrastructure-as-code, containers, and observability and monitoring of models and services in production.
Good cloud platform knowledge, with AWS preferred and Azure or GCP valued. Experience with a modern data and AI platform such as Databricks is a bonus.
Practical model evaluation experience: building benchmarks and tests that measure quality, cost, latency and safety, and producing results others can rely on.
Hands-on experience serving or running open weight models, including self-hosting and scaling inference. Fine-tuning or adapting models to proprietary data and workflows is ideal.
Working knowledge of API security, identity and access control, enough to implement MCP authentication and authorization correctly and work well with security teams.
Awareness of the governance context around enterprise AI, such as model provenance and licensing, data sovereignty and financial-services regulation.
A clear communicator and collaborator who works well with senior engineers, hands-on peers and specialist governance and security teams.
Deep curiosity about technology, especially AI, and the ability to show you learn quickly beyond your own specialty.
A self-starter who, once given clear direction, can reach a solution or deliverable without close supervision.
Requirements
Strong knowledge of modern AI: LLMs and foundation models, agents, and the tools and protocols, such as MCP, that connect them to enterprise systems.
Good software engineering skills. You write well-tested, production-quality code and build services other teams can depend on, not one-off scripts or prototypes.
Solid MLOps experience: CI/CD, infrastructure-as-code, containers, and observability and monitoring of models and services in production.
Practical model evaluation experience: building benchmarks and tests that measure quality, cost, latency and safety, and producing results others can rely on.
Working knowledge of API security, identity and access control, enough to implement MCP authentication and authorization correctly and work well with security teams.
Good cloud platform knowledge, with AWS preferred and Azure or GCP valued.
Hands-on experience serving or running open weight models, including self-hosting and scaling inference.
Preferred: experience with Databricks, fine-tuning or adapting models to proprietary data and workflows, and awareness of model provenance and licensing, data sovereignty and financial-services regulation.
Preferred: a clear communicator and collaborator, deep curiosity about technology, and the ability to show fast learning beyond your own specialty.
Preferred: a self-starter who, once given clear direction, can reach a solution or deliverable without close supervision.
All employees are responsible for information security: follow security policies and practices, protect the confidentiality and integrity of information you access, report any suspected breach, and complete mandatory security training.
Conditions
Hybrid or office setup, place and schedule not specified in the posting.