AI Agent Developer

Role Overview

The team is seeking a hands-on AI Agent Developer who can design, build, and iterate on data-driven AI agents using Microsoft technologies such as Microsoft Fabric, Azure, and Copilot Studio. The role requires strong engineering fundamentals, fast iteration, the ability to anticipate issues, and effective collaboration with both technical and nontechnical stakeholders.

• Build and deploy AI data agents connected to enterprise datasets and semantic models in Microsoft Fabric.
• Configure prompts, grounding strategies, exemplars, and agent instructions.
• Implement multiagent architectures, including parent/child routing, delegation, and fallback patterns.

• Write clear system messages, input/output formats, and guardrails.
• Evaluate prompt behaviour for reliability, correctness, retrieval quality, and output safety.
• Instrument evaluation loops and continuously improve models based on telemetry and feedback.

• Integrate agents with Fabric Lakehouse/Warehouse datasets and semantic models.
• Work with data engineers to define data requirements and ensure proper schema alignment.
• Map user intents to the correct data fields and grounding sources.

• Build supporting utilities and APIs using Python, SQL, and Azure services.
• Integrate agents with retrieval layers such as Cognitive Search, vector indexes, or API endpoints.

• Move quickly from MVP v1 vNext with minimal hand holding.
• Detect and diagnose issues such as data gaps, grounding mismatches, prompt failures, or routing errors.
• Recommend fixes and drive continuous improvements.

• Communicate technical reasoning clearly to engineers, architects, PMs, and nontechnical partners.
• Break work into clear phases, milestones, dependencies, and delivery plans.

• Proven experience building AI agents, copilots, or conversational AI applications.
• Strong engineering skills in SQL
  , Python
  , and Azure services.
• Experience connecting agents to semantic models, BI datasets, or Fabric-based data sources.
• Ability to design and operate multiagent systems (parent/child agents, routing logic, state handling).
• Strong problem-solving skills and ability to iterate quickly with limited guidance.
• Clear communication skills across technical and nontechnical audiences.

• Experience with Copilot Studio
  , agent orchestration frameworks, or Fabric Data Agents.
• Familiarity with retrieval-augmented generation (RAG), vector search, metadata enrichment, and evaluation pipelines.
• Understanding of enterprise data governance, privacy practices, and data stewardship.
• Experience creating telemetry dashboards or analytics to measure agent performance and quality.