Senior Staff AI Enablement Engineer

About the role

SoFi is scaling how employees and AI agents safely reach the tools they need to do their jobs. We run a centralized MCP (Model Context Protocol) gateway that lets approved AI clients (Claude, Cursor, Windsurf, internal agents) connect to enterprise systems such as Slack, Google Workspace, Jira/Confluence, GitLab, Datadog, Snowflake, Glean, and ServiceNow, with SSO/OAuth brokering, policy enforcement, and full auditability.

As our Sr. Staff AI Enablement Engineer, you'll own the strategy and day-to-day operation of this platform and the customer-facing interfaces employees use to get AI tool access. You'll turn a security-approved control plane into a fast, self-service path to AI productivity across the company. You'll partner closely with Product Security, IAM, Engineering, and business leaders.

What you'll do

• Own the MCP gateway platform end to end. This covers the catalog and registry of approved MCP servers, onboarding of new connectors, OAuth scope configuration, tool-level enablement and disablement, and lifecycle and hygiene.
• Design the customer experience for AI tool access. Build intake, request, and provisioning flows (Compass/ServiceNow, Slack, Okta), documentation, and self-service paths that cut time-to-access from weeks to hours.
• Partner with Product Security and IAM on MCP approval, threat modeling, least-privilege scopes, and incident response. Turn security requirements into operational guardrails that don't block adoption.
• Run operations for the platform. This includes L1/L2 support models, runbooks, SLAs, monitoring and observability of tool usage, and vendor management (e.g., Cequence and SaaS connector vendors).
• Drive adoption and measure impact. Define and report adoption, usage, and value metrics by function to IT and SLG leadership, and identify the next high-leverage integrations.
• Automate access governance. Integrate the gateway with joiner/mover/leaver processes, access reviews, and audit evidence for SOX and regulatory requirements.
• Set technical direction. Evaluate emerging agent and tool-access standards (MCP, A2A, agent identity), build vs. buy, and the roadmap for internal and customer-facing AI interfaces.
• Mentor and influence. Guide IC1–IC4 engineers and analysts, and influence function-level decisions with senior ICs and SLG.

What you'll need

• Education & Minimum Experience Requirements: Bachelor’s Degree + 4+ Years of Experience OR Master’s Degree + 3+ Years of Experience in software engineering or a related field.
• Track record of owning complex, companywide programs or platforms and independently handling their planning, execution, and resourcing.
• Hands-on experience with LLM and agent tooling: MCP servers and clients, API gateways, and AI coding assistants such as Claude Code, Cursor, or Copilot.
• Deep knowledge of identity and access: SSO/SAML, OAuth 2.0/OIDC, SCIM, Okta, and least-privilege and RBAC design.
• Strong integration and automation skills. That means REST/GraphQL APIs, scripting (Python, TypeScript, or similar), iPaaS (Workato or similar), and ServiceNow workflows.
• Experience operating in a regulated environment (financial services preferred), including audit, data retention, and change management.
• Ability to translate between security, engineering, and business stakeholders, and to influence without authority.

Nice to have

• Experience building or running an AI gateway, API gateway, or developer platform (e.g., Cequence, Kong, Apigee).
• Experience building Slack bots, deflection bots, or agentic workflows for IT or employee support.
• Familiarity with AI governance frameworks (NIST AI RMF) and LLM security risks (OWASP LLM Top 10, prompt injection, tool poisoning).