Director of IT

This job combines two things that usually sit with different people: owning products and building them. One product is a real-time tracking and communications app that hundreds of thousands of families and their school districts rely on, where a wrong notification reaches a parent waiting at a bus stop. The other is the technology behind a service for students with special transportation needs, students experiencing homelessness, and out-of-district students, delivered through contracted local transportation providers.
Each product has a Business Product Owner who owns its business outcomes, priorities, and commercial decisions. You are their partner on the IT side, owning how those priorities become software within the lines our product management operating model sets. It gets built in an agentic delivery framework owned by a senior technical team. You own its product side: the specs, acceptance criteria, and evals that go in, and the call on whether the result met the business need.
You also build. In traditional product management, requirements get written, handed to engineering, and interpreted somewhere along the way. Here, a conversation with a stakeholder can become working software in days, and the person who had the conversation is often the one directing the agents that build it. You should already be building with agents, even if it is not yet how you do all of your work. In this role, building stops being something you do alongside product management and becomes how you do it.
You will start with a small team, so expect to lead and build at the same time. The people you lead may come from more traditional roles, and helping them make the same shift is part of the job. You will also own the budget, vendor relationships, and contracts behind this work, AI tooling included.
How You'll Work
AI runs through your entire workflow, from discovery and specs to the build, tests, documentation, and reporting. We are looking for someone already working this way who can show what it has produced, not someone with a thesis on where AI is going.
Requirements here usually arrive as a conversation rather than a ticket, and you will often be the one having it. You will build inside a framework someone else runs, which means understanding a rule's reasoning before deciding it is in your way. You will be measured on whether what ships is correct and whether the business got what it actually needed, not on volume.
A typical week may include:
Sitting with district staff, families, or the team that coordinates our transportation providers, then deciding with a Business Product Owner whether a problem belongs on the roadmap.
Turning a conversation like that into a working prototype and using it to agree on requirements before the full build starts.
Reviewing work with a team member who is still learning to build with agents, and explaining the reasoning behind your feedback.
Key ResponsibilitiesOwn the products
Co-own each product's strategy, roadmap, and lifecycle with its Business Product Owner.
Manage the backlog and sequence work against business goals and the dependencies between the two products, not just the request queue.
Develop business cases and set each product's metrics, from family adoption and engagement to district retention and safe, on-time service. Partner with the business on commercialization and pricing.
Decide where AI belongs inside the products, such as how families get answers or how trip exceptions get handled, and define the evals that judge whether it works.
Build with agents
Direct AI coding agents to build applications, workflows, and AI-enabled tools from first prototype through production, and stay with what you ship after it is live.
Read all agent output critically, including your own. Reject work that only looks finished, and use tests, evals, and the framework's gates to prove behavior rather than assert it.
Own the product side of agentic delivery
Own how work enters the framework, from the prototypes used to agree on requirements to the specs, acceptance criteria, and evals that define done, and how delivered work is judged against the business need.
Define the data, privacy, and safety requirements for your products, with the highest bar for students whose information needs the most protection, and work with the framework's owners so the system enforces them.
Bring the senior technical team specific evidence on where the framework slows product work down, and make the case for tools and techniques that earn their place.
Work directly with the business
Elicit requirements from people who describe outcomes, not specifications, and ask the questions that surface the real constraint: who uses it, how often, and what happens when it fails.
Ground decisions in a deep understanding of your users, from families following a student's ride to district staff, operations teams, and local providers, and test designs with them early.
Partner with business leaders on how technology can serve their goals, lead cross-department work, and keep leadership current on what was decided and why.
Lead and grow the team
Develop the people you lead into agentic roles along a clear path: requirements, then agent-ready specs and evals, then prototypes, then small changes shipped to production.
Teach through real work, with reviews that build judgment, not just fix mechanics. Recruit and retain the people who will grow this capability as the team expands.
Manage the investment
Own the budget for your products and team, including AI tooling and model usage, and find savings where agents can now do work that once required a vendor or a larger team.
Source and manage vendors, negotiate contracts, and hold vendors to their commitments.
Required Qualifications
Hands-on experience building working software with AI coding agents (Claude Code, Codex, Cursor, or successors), beyond using them as autocomplete. Show us what you have built and explain the decisions behind it.
10+ years across product management, software delivery, and technology leadership, with a history of owning product strategy and delivering measurable results.
Experience leading products alongside business owners under clear decision rights.
Experience turning vague, verbally described needs into the specs, acceptance criteria, and evals that people and agents build from, and an understanding of why a passing test is not proof that a change is correct.
Ability to read code you did not write, well enough to catch an agent's mistakes and explain them, in more than one language or platform.
Solid grasp of software architecture and data principles, from service boundaries and where state lives to data quality and the handling of sensitive data.
Experience hiring and developing people, including helping someone grow into work their p