MCP Integration Engineer
Description
Who we are
We are looking for a MCP & Integrations Developer to design, build, test, deploy, and maintain tools and integrations that connect enterprise services with AI agents and platform capabilities. The role focuses on reliable backend integration engineering, secure API consumption, identity and permission handling, and end-to-end production readiness.
The engineer will work autonomously across the full integration lifecycle, from technical analysis and implementation to automated testing, observability, incident diagnosis, and ongoing maintenance. Deep expertise in every target connector is not required; the key expectation is the ability to develop and support new integrations quickly and safely using consistent engineering patterns.
What you'll be doing
- Design, develop and maintain MCP servers, agent tools, backend services and integration components using Python, FastAPI and Pydantic
- Integrate external and enterprise services through REST APIs, SDKs, webhooks and other supported interfaces
- Implement secure authentication and authorization flows using OAuth 2.0 and OIDC, including token lifecycle, scopes, permissions, credentials and secret handling
- Develop reusable integration patterns for tools consumed by AI agents, including clear schemas, validation, error handling and predictable responses
- Build and maintain automated unit and integration tests with pytest and validate integrations end to end across development and production-like environments
- Deploy and operate integration services on Azure using services such as Azure Functions, Cosmos DB and Key Vault
- Contribute to CI/CD pipelines, environment configuration, release automation and controlled deployment practices
- Instrument services with logs, metrics, traces, and actionable diagnostics to support monitoring and troubleshooting
- Diagnose incidents across API, authentication, permissions, data, network, and platform layers; perform root-cause analysis and implement sustainable fixes
- Document APIs, MCP tools, authentication requirements, configuration, operational procedures, known limitations and support guidance
- Collaborate with Platform, Security, IAM and product teams when integrations require shared capabilities, permissions, or infrastructure changes
What you'll bring along
- Strong hands-on Python development experience and practical knowledge of FastAPI and Pydantic
- Experience building REST APIs and production-grade backend integration services
- Knowledge of Model Context Protocol (MCP) or equivalent experience building APIs, tools, or function interfaces for AI agents
- Practical experience with OAuth 2.0/OIDC, authentication and authorization flows, token management, permission models, credentials and secrets
- Experience with Git, code review practices and automated testing using pytest
- Experience with Azure Functions, Cosmos DB, Key Vault, CI/CD pipelines and cloud-based service configuration
- Ability to add observability and troubleshoot distributed integration flows using logs, metrics, traces, and correlation information
- Understanding of secure software-development practices for APIs and enterprise integrations
- Experience with Microsoft Graph integrations, including Outlook, SharePoint, or Teams, is highly valued
- Experience integrating Jira, Confluence, GitLab, Azure DevOps, or similar enterprise platforms is an advantage
- Proven experience delivering and maintaining backend APIs or enterprise integrations in production
- Demonstrable ability to take an integration from technical design through implementation, testing, deployment, monitoring and support
- Ability to work independently, manage technical ambiguity, and learn unfamiliar APIs or connector domains quickly
- Strong problem-solving skills and a disciplined approach to debugging, testing, documentation and maintainability
- Professional working proficiency in English to understand technical documentation and collaborate effectively with international teams
- Candidates are not expected to be experts in every connector; they must be able to develop, test, operate and maintain integrations autonomously