AI DevOps & Infrastructure
AI Agent
Nodus: AI agent using custom, secure MCP server on scalable infrastructure
“Our experience working with Tech 42 was excellent. Their team helped us design and implement a secure MCP (Model Context Protocol) server environment using a Docker-in-Docker approach that enables safe, isolated Python code execution for data analysis. They delivered a fully containerized solution pushed to our ECR repository and provided clear guidance and support during integration with our existing agent architecture. The collaboration was smooth, communication was consistent, and their technical expertise allowed us to achieve our goal quickly and confidently. We look forward to working with Tech 42 in the new year as we expand our AI workflows.”
Daniel Sanchez
Director of Engineering
Project summary
Nodus needed to enable AI-driven data analysis through Amazon Bedrock agents while securing AI-generated Python code. Tech 42 designed and built a containerized MCP server using Docker-in-Docker for isolated, secure code execution. The solution integrates with Bedrock agents through the Model Context Protocol, supports core data science libraries, and handles data access and artifact storage through Amazon S3. Users can now run sophisticated data manipulation and visualization tasks without manual coding. The fully containerized solution was deployed to ECR and documented for seamless integration with the existing agent architecture. The result is a secure, scalable way for AI agents to execute analysis workloads on demand.
