This is a fully remote position, open to applicants in California, +1 more state.

📋 Description

• Oversee the MCP fleet operations, which include shipping, securing, and applying patches to self-hosted and managed MCP servers for business applications and databases, as well as managing the MCP registry.

• Take charge of the deployment patterns, guardrails, observability, and hosting infrastructure of the agent platform.

• Co-manage the Bedrock runtime and gateway, focusing on model access, cost attribution, content-filtering guardrails, model pinning, and quota management through an AWS/Okta-SSO-secured gateway.

• Develop the Bedrock-to-Snowflake pipeline, incorporating identity matching, usage tracking, and threshold alerting functionalities.

• Manage MDM configuration, set up desktop clients, and handle credentials and authentication processes for non-technical personnel.

• Create AI automation solutions across Claude desktop, Slack, and Google Workspace.

• Direct workflows based on data sensitivity, enforce least privilege principles and auditability, and isolate untested models from production data.

⛳️ Requirements

• A minimum of 6 years in software engineering, platform engineering, or DevOps, with ownership of a production component relied upon by others.

• Practical experience with AWS — including IAM, ECS or Lambda, CloudWatch, and a solid understanding of VPC. This is essential, not optional.

• Comprehensive full-stack skills — encompassing backend services, APIs, CI/CD processes, and containerized deployments.

• Proficiency in writing an MCP server or API integration from the ground up, rather than merely connecting a managed service.

• Direct experience with LLMs, LLM APIs, prompt engineering, and agent development; must have successfully delivered a project involving these technologies.

• Capability to interpret an ambiguously defined task and identify the undocumented components necessary to complete it.