In this project, I designed and introduced an MCP-based integration layer for the development team, connecting everyday engineering tools directly with the AI code assistant inside JetBrains IDEs.
The goal was to create a practical AI-enhanced developer environment where the team could work faster without constantly switching between systems. The integration covered Jira search, reading and creating tasks, access to multiple PostgreSQL databases, multiple Elasticsearch clusters, server monitoring tools, Graylog, and Bitbucket. Part of the work also included preparing prompts, MCP server setup, and running a training session for the team so the solution could be used effectively in day-to-day development.
A key part of the implementation was standardizing the MCP configuration so it could be stored in the project under .ai/mcp/mcp.json, making the setup more maintainable and easier to share across the team. The documentation also covered the JetBrains AI Assistant MCP setup, Atlassian integration flow, PostgreSQL integration, troubleshooting, and security practices such as using read-only database accounts and being careful with sensitive data in prompts.
From a leadership perspective, my role as Tech Lead was not only technical delivery, but also building the internal AI foundation around the team: MCP servers, prompt conventions, onboarding guidance, and training. One important lesson from this initiative was that building the infrastructure is only one part of adoption. To get the full value from AI enablement at team level, it is also important to define and consistently follow a lightweight tracking method for usage, outcomes, and knowledge sharing.
This project helped move AI support in the IDE from isolated experiments to a more structured engineering capability embedded into the team’s daily workflow.
My role
Tech Lead
Scope
- MCP architecture and setup
- JetBrains AI Assistant integration
- Jira integration for search and task operations
- PostgreSQL integration
- Connections to Elasticsearch, Bitbucket, Graylog, and monitoring tools
- Prompt preparation and internal usage guidance
- Team training session
Key outcomes
- Reduced context switching between IDE and external systems
- Enabled AI-assisted access to engineering and operational tools
- Standardized MCP configuration inside the project structure
- Improved team readiness to use AI tooling in real development work
Technologies
JetBrains AI Assistant, MCP, Jira, PostgreSQL, Elasticsearch, Bitbucket, Graylog