AI Tooling & MCP Servers
I build the plumbing that lets AI agents work with real internal systems without exposing sensitive data. At Sharjah Digital Department I architected a suite of Model Context Protocol (MCP) servers and engineered agents, retrieval-augmented generation (RAG) pipelines and guardrails on the AIRIA orchestration platform.
What I deliver
-
MCP servers
Structured, permission-aware interfaces to internal systems: an Azure DevOps Server MCP server (projects, repositories, pipelines, work items, build history), an Active Directory MCP server for identity and group lookups, and a localisation MCP server.
-
Localisation automation
The TransLynk MCP server scans application source for UI strings, generates English and Arabic translations, diffs them against existing translations and upserts only net-new entries.
-
Agents, RAG and guardrails
AI agents, RAG pipelines, security guardrails and intent-based tool routing for policy-compliant automation.
-
Internal AI assistants
A context-aware assistant that answers from governance, HR and IT policy content and the service catalogue using RAG.
-
AI-assisted engineering
GitHub Copilot and prompt engineering to accelerate day-to-day development workflows.
Selected work
-
AI Tooling Suite — MCP servers & AIRIA orchestration
.NET, MCP server SDK, Azure DevOps REST API, LDAP / Active Directory, AIRIA platform.
-
OnePortal AI Assistant
A context-aware chatbot inside OnePortal using AIRIA RAG pipelines over policy content and the service catalogue.
Technologies
- Model Context Protocol (MCP)
- .NET
- AIRIA
- RAG
- Azure DevOps REST API
- LDAP / Active Directory
- GitHub Copilot