About FacilityOS

FacilityOS is a fast-growing company redefining how facilities operate—bringing safety, security, and daily operations into one unified platform used by organizations around the world.

As we continue to scale globally, we’re building a team of driven, curious people who want to make an impact. You’ll be part of a dynamic, collaborative culture where individuals are trusted to take ownership, solve meaningful problems, and grow in their careers. Our team comes together in-office twice a week to connect, collaborate, and build momentum.

If you’re looking to do your best work alongside a great team in a high-growth environment, FacilityOS is the place to build your career.

About the Role

The AI Engineer reports to the VP of Engineering and serves as FacilityOS's hands-on technical anchor for AI agent development. This role sits within the central engineering function that acts as an integrator and enabler for the AI agents that individual departments design and own, building the pipelines, frameworks, and integrations that make department-defined agents production-ready, secure, and scalable.

Department leaders own the discovery, tooling choices, behaviour, and outcomes of their own agents. The AI Engineer's role is to build and maintain the underlying technical architecture, RAG pipelines, agent orchestration platforms, and API integrations into business systems such as Microsoft 365, email, and Teams that supports that department-level ownership, not to run agents on departments' behalf.

Key Responsibilities

Agent Architecture & Development

Design, build, and maintain the technical architecture underpinning FacilityOS's AI agents, including RAG pipelines, agent frameworks, and orchestration layers

Integrate department-selected AI tools and platforms into a consistent, secure, and supportable technical stack

Own the organization's approved AI/agent tooling stack in partnership with the VP of Engineering and CTO, evaluating new tools and frameworks as they emerge

Evaluate, select, and operate agent orchestration platforms (e.g., OpenClaw or comparable agent orchestration/runtime frameworks) that coordinate multi-step agent workflows, tool calls, and task hand-offs

Business Systems Integration

Connect agents into the core business systems departments already work in - Microsoft 365, Outlook/email, and Teams, so agents can read, draft, notify, and act within existing workflows rather than requiring new interfaces

Build and maintain secure authentication and permissioning patterns (e.g., Microsoft Graph API, OAuth) for agents acting inside business systems on a user or department's behalf

Establish reusable integration patterns for business-system connectivity so each department's agents don't require one-off custom builds

Agent Evaluation & Quality

Build and own the evaluation infrastructure that determines whether an agent actually works — eval harnesses, regression testing for prompts and workflows, and quality/accuracy metrics

Establish guardrails and mitigations for hallucination, unsafe outputs, and failure modes, with clear quality gates an agent must pass before deployment

Provide departments with reusable evaluation patterns so they can measure and improve their own agents' quality over time

Cross-Functional Enablement

Partner with the Data Engineer to ensure agents have reliable, well-structured data to orchestrate against

Act as the technical resource within the AI Tech/Integration SWAT, unblocking department-level technical issues without taking ownership of department agent decisions

Translate department managers' agent requirements into working technical implementations

Compliance & Deployment Readiness

Build agents to meet FacilityOS's ISO27001 and SOC2 Type 2 compliance requirements from the ground up

Prepare technical documentation and evidence required for VP Infosec's mandatory pre-deployment review of every agent build

Maintain security, monitoring, and logging standards across all deployed agents

Program Support

Provide technical input to the AI Program Lead's portfolio planning, flagging technical risk and feasibility early