As part of Diabsolut’s continued growth, we are expanding our team and seeking someone who has experience building an enterprise-grade agentic platform that orchestrates multiple specialized AI agents across delivery, operations, and client engagement. The platform uses MuleSoft Agent Fabric (MAF) as the central broker, wiring together MCP data servers, Azure AI Foundry reasoning engines, and Salesforce Agentforce agents into a unified, governed network.

We are looking for an AI Integration Engineer to join as a core builder. You will work directly with the Lead Architect to design, develop, and operate the agent network — from MAF broker configuration and MCP server development to Azure AI agent wiring, interface adapters (Slack, Teams, custom UI), and production deployment. This is a hands-on engineering role where you will ship working code every week.

This role is unique: you won’t just build integrations — you’ll build the AI agents that use them, and the infrastructure that runs them securely at scale.

What You’ll Build

Agentic Orchestration — MuleSoft Agent Fabric (Primary ~40%)

• Build and maintain the MAF broker — YAML-driven agent orchestration, intent routing, session management, and persona profiles
• Develop and publish MCP servers (Model Context Protocol) — reusable data connectors for Salesforce/Certinia, Azure DevOps, SharePoint, and internal knowledge indexes
• Implement DataWeave transformations for API response mapping, data normalization, and payload shaping across heterogeneous sources
• Configure Flex Gateway policies — rate limiting, PII detection, Cedar-based RBAC, and response filtering Deploy and manage applications on CloudHub 2.0 (Private Space, monitoring, auto-scaling) and/or Runtime Fabric
• Register and version MCP servers in Anypoint Exchange for governance and discoverability
• Secure and rotate integration credentials — Azure Key Vault, Salesforce Connected Apps, Entra App registrations

AI Agent Development — Azure, Agentforce & A2A (~35%)

• Build reasoning agents on Azure AI Foundry — system prompts, knowledge base configuration, GPT-4o tuning, and tool registration Develop A2A adapters (Agent-to-Agent) using Python Azure Functions — JSON-RPC protocol translation across agent boundaries
• Implement multimodal capabilities — GPT-4o Vision for images, Azure Document Intelligence for PDFs and structured documents
• Write and refine agent system instructions — prompt engineering for domain accuracy, guardrails, and response quality
• Build Agentforce agents with Apex invocable actions and Einstein AI integrations for CRM-native workflows
• Integrate Microsoft Copilot Studio or Power Automate where M365-native orchestration is required

Interface Adapters — Slack, Teams & Custom UI (~10%)

• Build and maintain the Slack adapter — Socket Mode / HTTP Events API, Block Kit rich output, file export (PPTX, Excel) o Build Microsoft Teams integration — Bot Framework, Adaptive Cards, message extensions for agent-delivered outputs
• Develop lightweight custom UI surfaces (React or equivalent) for use cases that require richer interaction than chat
• Design and implement consistent output formatting across channels — text, tables, charts, and generated artifacts

Production Operations, Security & CI/CD (~15%)

• Implement production monitoring, alerting, and audit logging across the broker, MCP servers, and agent endpoints
• Design and enforce role-based access control — persona routing profiles in MAF, Entra ID groups, Salesforce Permission Sets
• Maintain CI/CD pipelines for MuleSoft and Azure Container Apps deployments (Azure DevOps pipelines) Run load tests, establish performance baselines, and optimize for latency and throughput across the agent network
• Coordinate with IT on Entra App approvals, SharePoint permissions, network security, and compliance requirements
• Deploy across environments (dev / staging / production) with environment-specific config management and rollback capability

Technology Stack

You will work across these platforms daily:

-**Orchestration:**MuleSoft Agent Fabric, Anypoint Platform, CloudHub 2.0, Flex Gateway -**Data Layer (MCP):**Salesforce Connector, HTTP Connector, DataWeave, RAML / OAS, Anypoint Exchange AI / Agents: Azure AI Foundry, Azure OpenAI (GPT-4o), Azure Document Intelligence, AI Search -**Agent Protocols:**A2A (JSON-RPC), MCP (Model Context Protocol), Azure Functions (Python) -**Salesforce:**Certinia PSA, SOQL, Connected Apps, Agentforce, Einstein, Apex -**Interfaces:**Slack API (Socket Mode / Block Kit), Microsoft Teams (Bot Framework / Adaptive Cards), Custom React UI Security: Azure Key Vault, Entra ID, OAuth 2.0, Cedar policies, Flex Gateway RBAC -**DevOps / CI-CD:**Azure DevOps (ADO Pipelines), Git, Python, Node.js, Azure Container Apps

Team and How We Work You’ll join a small, high-output team building something new.

• The platform is in active development — the POC is live, the architecture is proven, and we’re scaling to production. You’ll work directly with the Lead Architect daily and collaborate with IT, MuleSoft, and Salesforce teams as needed. -**Ship weekly —**short iterations, working software over documentation
• Architecture-first — every agent follows the same broker + MCP + reasoning pattern (governed by ADRs)
• AI-assisted development — we use Claude Code and AI tooling extensively for code generation, testing, and architecture
• Ownership mindset — you own what you build, from development through production monitoring Remote-first — async communication with scheduled syncs (weekly SteerCos, team standups)

Required Skills & Qualifications:

Must Have

• 3+ years MuleSoft development — Mule Runtime, DataWeave, connectors (Salesforce, HTTP, Database), API-led design
• Experience deploying to CloudHub 2.0 or Runtime Fabric — configuration, monitoring, troubleshooting Strong REST API integration skills — OAuth 2.0 (client credentials, JWT bearer), Microsoft Graph API, Azure DevOps REST
• Python proficiency — scripting, Azure Functions, API development (FastAPI or Flask)
• Experience with Azure cloud services — Key Vault, Container Apps, Entra ID, or equivalent cloud platform
• Understanding of LLM-based systems — prompt engineering, token management, model API interaction (Azure OpenAI or equivalent)
• Git workflow, CI/CD fundamentals, and comfort working in a fast-moving team with short iteration cycles Strong Differentiators
• MuleSoft Agent Fabric — agent network YAML, A2A connectors, MCP server development (rare but ideal) Multi-agent architecture patterns — orchestration, A2A protocol, agent-to-agent handoff and state management
• Model Context Protocol (MCP) — server development, tool registration, resource exposure to LLMs Salesforce platform — SOQL, Connected Apps, Permission Sets, Apex basics, Agentforce / Einstein Azure AI services — AI Foundry, Azure OpenAI, Document Intelligence, AI Search, Speech-to-Text Slack API — Socket Mode, Events API, Block Kit, interactive components
• Microsoft Teams / Bot Framework — Adaptive Cards, message extensions, Bot Channels Registration Certinia PSA (FinancialForce) data model — projects, resources, milestones, billing
• MuleSoft certifications (MCD, MCIA) or Salesforce certifications (Platform Developer, Integration Architect)

About Diabsolut

At Diabsolut, we consult on and implement innovative Salesforce and Certinia solutions to help organizations across diverse industries achieve success and deliver exceptional outcomes.

As a Salesforce Summit Consulting Partner and Certinia Premier Partner, we provide specialized expertise in Education Cloud, Revenue Cloud, Field Service, Certinia Professional Services Cloud, Customer Success Cloud, Accounting, and more.

Our mission is to empower clients with not only adaptable solutions but also the knowledge to navigate future challenges confidently. We pride ourselves on building long-lasting relationships alongside robust, scalable solutions. We are heavy users of AI within our delivery to improve both time to value and likelihood of our customers outcomes.