Description:

About Us

Preferred Business Systems (PBS) has been making technology simple for small and midsize businesses since 1997. From our headquarters in Whippany, New Jersey, we deliver managed IT, cybersecurity, cloud, business communications, and office technology to organizations across New Jersey, New York, Pennsylvania, Connecticut, and the rest of the U.S.

Our clients span healthcare, law, logistics, warehousing, manufacturing, construction, hospitality, and financial services. Our growing AI and Automation team builds AI-powered agents that take repetitive and complex work off their plates, from executive assistance and recruiting to warehouse dispatching and customer communications.

The Role

As an AI Agent Developer, you'll design, build, integrate, and maintain AI agents that streamline real business operations. You'll work within the architecture and standards set by technical leadership, owning the implementation details that turn an approved design into a working solution.

You'll partner with Business Strategists, who lead customer relationships and define requirements. Your job is to turn those requirements into automations that are functional, reliable, and secure.

Prior AI agent experience is a plus but not required. If you have a strong foundation in software development, APIs, and automation, we'll help you grow into the AI side.

What You'll Do

Build and integrate AI agents

• Develop AI agents using Microsoft Copilot Studio, Power Automate, Azure services, and custom code.
• Build custom applications and integrations in Python, TypeScript, or similar languages.
• Connect agents to Microsoft 365, Microsoft Graph, third-party APIs, databases, and customer business applications.
• Implement LLM-powered workflows, tool calling, conditional logic, and human approval steps.
• Manage LLM usage costs and performance as part of every build.
• Create reusable components that speed up future implementations.

Work with customers

• Join Business Strategists in technical discovery to identify integration needs and system limitations.
• Support customer demos, solution testing, and troubleshooting.
• Keep internal stakeholders informed on progress, challenges, and risks.

Test, deploy, and improve

• Write and run functional, integration, and error-handling tests.
• Evaluate LLM output quality, including consistency, hallucinations, and edge cases.
• Validate accuracy, permissions, and reliability before deployment.
• Deploy approved solutions into customer environments.
• Resolve escalated defects and improve agents based on performance data and customer feedback.

Build securely and responsibly

• Implement secure authentication, authorization, and least-privilege access.
• Protect customer data and credentials using approved practices.
• Identify risks such as sensitive data exposure, prompt injection, and unauthorized actions.
• Build in human approval checkpoints and audit logging.
• Participate in security reviews and remediate findings.

Document and hand off

• Maintain documentation for agents, integrations, configurations, and deployments.
• Write troubleshooting guides and support runbooks.
• Define monitoring, alerting, and logging requirements.
• Train our support team so every deployed agent is supportable long-term.

Requirements:

What You Bring

• 2+ years of professional software or automation development experience.
• Proficiency in Python, TypeScript, or a comparable language, using Git and standard development practices.
• Experience building and consuming REST APIs and integrating with third-party systems.
• Solid understanding of OAuth 2.0, authorization, and secure credential handling.
• Familiarity with Microsoft 365 and Power Automate (or a comparable workflow platform).
• Strong debugging and problem-solving skills.
• Clear communication with both technical and nontechnical people.

Nice to Have

• Experience building AI agents or LLM-powered applications.
• Familiarity with Microsoft Copilot Studio, Microsoft Graph API, and Azure services.
• Understanding of LLM APIs, function calling, and agent orchestration.
• Familiarity with retrieval-augmented generation (RAG) and vector databases.
• Experience with AI evaluation, monitoring, and observability.
• SQL and data integration experience.
• Experience in an MSP or multi-customer environment.
• Familiarity with CI/CD pipelines and cloud deployments.