Microsoft AI Engineer (Copilot/Power Platform Agent Developer) Full Time

Job Title: Microsoft AI Engineer (Copilot/Power Platform Agent Developer)

AI Agent Developer / Copilot Studio Specialist

Location: Remote (Occasional travel to Hyderabad)

Shift: US timezone overlap (IST evenings)

Experience: 2–5 years in Microsoft Power Platform / AI / Automation

The Role

We are looking for a hands-on AI Agent Developer who can build, deploy, and drive adoption of Microsoft Copilot agents for enterprise clients. This is not an architecture or strategy role — this is a delivery role where you sit with business users, understand their workflows, and build agents that automate their repetitive tasks.

You will be deployed on client engagements where the end users are non-technical operations, logistics, sales, and finance teams. Your job is to translate their manual processes into working Copilot agents — from requirements gathering to production deployment to user training.

Think of this as a hybrid of consultant and builder: you run a requirements workshop in the morning and build the agent in the afternoon.

What You’ll Do

• Meet with business users (individually and in small teams) to understand their daily workflows, identify repetitive tasks, and gather requirements for automation
• Build agents in Microsoft Copilot Studio — configure prompts, design conversation flows, connect knowledge sources, set up triggers, and publish to Microsoft Teams
• Create Power Automate flows that integrate with Copilot agents — SharePoint triggers, email parsing, data extraction, document generation, and API connectors
• Connect agents to enterprise data sources — SharePoint, Oracle, Salesforce, and third-party APIs. Configure custom connectors where out-of-the-box options don’t exist
• Test, iterate, and refine agents based on user feedback. Debug prompt issues, handle edge cases, and optimize output quality across different LLM models
• Drive user adoption — train non-technical users on how to use, maintain, and iterate on their agents. Create documentation and quick-start guides
• Document and report — maintain a backlog of use cases, track agent performance, and provide weekly status updates to stakeholders

Must-Have Skills

Copilot Studio

Hands-on experience building agents in Microsoft Copilot Studio — not just awareness, actual builds shipped to users. Topics, actions, knowledge sources, custom instructions, publishing.

Power Automate

Building flows with triggers (SharePoint, email, scheduled), connectors (HTTP, custom), conditions, loops, and error handling. Experience integrating flows with Copilot agents.

Microsoft 365

Working knowledge of SharePoint, Teams, Outlook, and OneDrive as integration surfaces. Understanding how data flows across the M365 ecosystem.

Prompt Engineering

Ability to craft, test, and refine prompts for GPT-4/4.5 and other models. Understanding of token limits, context windows, and output quality optimization.

Requirements Gathering

Can sit with non-technical business users and translate their manual workflows into structured agent requirements. This is as much a consulting skill as a

technical one.

API Integration

Connecting Copilot Studio to external data sources via REST APIs. Understanding of custom connectors, authentication (OAuth, API keys), and BYOK model configuration.

Communication

Excellent written and verbal English. Comfortable presenting to senior stakeholders and coaching non-technical users. US-facing client communication

experience preferred.

Strong-to-Have Skills

• OCR / Document Processing: Experience with image-to-text, screenshot parsing, or intelligent document processing use cases
• Salesforce Integration: Familiarity with Salesforce APIs, MCP servers, or data extraction from Salesforce environments
• Azure AI Services: Azure OpenAI, Cognitive Services, Azure AI Search, or Azure Bot Service
• Domain Experience: Background in logistics, supply chain, sales operations, or global trade — understanding the domain dramatically accelerates requirements gathering
• Training & Adoption: Experience coaching non-technical users to maintain and iterate on their own agents and automations post-deployment
• Multi-model Awareness: Hands-on experience with Claude, Gemini, or open-source models in addition to GPT. Understanding of model strengths and trade-offs for different task types

Example Use Cases You’d Work On

These are representative of the kinds of problems you’d solve:

• Rate card optimization agent: Takes source documents from multiple shipping carriers, compiles optimized rate tables by region and freight level, triggered by SharePoint file upload
• Deal evaluation agent: Checks international deal parameters — currency alignment, entity setup, vendor agreements, taxes/duties — across multiple data sources and provides recommendations
• Error resolution agent: OCRs screenshots of system errors, matches against known error documentation, provides step-by-step troubleshooting in plain language
• Shipping quote automation: Takes intake parameters (origin, destination, weight, freight type), matches against rate cards, generates automated shipping quotes