Agentic AI Developer

Role Summary

The AI Agent Developer / Engineer builds and deploys enterprise AI agents across business domains, from design through to secure production release. In addition to core build responsibilities, this role owns security and risk remediation for agents already in production, and partners with the Red Teaming/AI Security Architect to red-team agents while they are still being developed — so that security issues are caught and fixed before, not after, go-live.

Key Responsibilities

Agent Build & Deployment

• Build and deploy enterprise AI agents for various business domains using Semantic Kernel, AutoGen, and Microsoft Agent Kit (Agent Builder).
• Develop reasoning, memory, and tool-calling capabilities within agents.
• Build connectors to enterprise systems (APIs, databases, SharePoint, and similar sources).
• Implement feedback loops and agent evaluation metrics to continuously improve agent quality and reliability.
• Support industry-specific agent pilots (e.g., Disaster Recovery, Healthcare, Manufacturing).

Security & Risk Remediation (Deployed Agents)

• Own technical remediation of security and risk issues identified in deployed agents, including excessive tool/data permissions, insecure connector configurations, exposed secrets, and unsafe prompt/tool-call patterns.
• Work directly with the AI Security/Remediation Engineer to action findings surfaced by security tooling and red team assessments, and confirm fixes before closure.
• Update agent architecture, configuration, or code to close identified gaps without breaking production functionality.

Red Teaming (Agents in Development)

• Partner with the Red Teaming/AI Security Architect to red-team agents while still in development, surfacing prompt injection paths, tool/function-call abuse, unsafe memory/context handling, and overly broad connector permissions before launch.
• Incorporate red team findings into the build process (secure defaults, guardrails, scoped tool access) so agents reach production review with fewer critical findings.
• Maintain a secure-build checklist for new agents based on recurring red team and remediation findings.

Required Qualifications

• 3+ years of software development experience in Python and/or Node.js, with hands-on experience building LLM-powered applications or agents.
• Practical experience with AI agent SDKs/frameworks such as Semantic Kernel, AutoGen, or Microsoft Agent Kit (Agent Builder).
• Experience building RAG pipelines, including vector search and embedding models.
• Experience implementing tool-calling, planning, and LLM API integration patterns.
• Working understanding of common AI/agent security risks (prompt injection, excessive tool permissions, insecure connector/credential handling) and willingness to remediate issues as part of the build cycle.

Preferred Qualifications

• Experience building connectors to enterprise systems (REST/GraphQL APIs, relational databases, SharePoint, or similar).
• Exposure to agent evaluation frameworks and feedback-loop/metrics design.
• Experience participating in or responding to red team exercises on AI/LLM systems.
• Domain exposure to regulated or high-stakes industries (healthcare, manufacturing, disaster recovery) where agent reliability and safety are critical.