Director of AI‑Driven Technical Program Management

Grant Thornton

Phoenix (AZ)

Hybrid

USD 184,000 - 260,000

Full time

4 days ago

Be an early applicant

Application generator

An application made for this job — a tailored resume and cover letter that speak straight to the posting.

Get past ATS filters

Job summary

Grant Thornton is seeking a Director of Technical Program Management to join the team. The role centers on turning strategic roadmaps into coherent, executable plans across Innovation, Platform, Delivery and Operations, with a strong emphasis on AI-enabled work and measurable outcomes.

The position requires extensive experience leading cross-functional teams, shipping software used by real users, and applying modern agile practices.

Qualifications

Twelve or more years in product engineering, with at least five leading technical program management or engineering delivery teams in a product organization.
A record of shipping software products that real users adopted, ideally including both internal-facing tools and externally used products.
Genuine technical depth: comfortable in architecture and design discussions, able to challenge an estimate or a dependency on the substance rather than the process.
Experience running large cross-functional programs spanning several engineering teams, product management and design.
Fluency in modern agile practice and flow metrics, used to improve delivery rather than to report activity.
Experience managing and developing program managers or scrum masters across distributed locations and time zones.
Working fluency with AI tooling in the delivery workflow, and judgment about where it helps and where it produces confident noise.
Experience program-managing AI or machine learning work, where quality is measured statistically, iteration is expected and a fixed-scope plan is the wrong instrument.

Responsibilities

Turn roadmap commitments into sequenced, resourced program plans with named owners, real dates and an explicit definition of the user outcome being delivered.
Run the programs that span multiple squads and pillars, from validated concept through build to a product increment in front of internal or external users.
Map dependencies across Innovation, Platform, Delivery and Operations early enough to be designed around rather than escalated late.
Drive scope and sequencing trade-offs with product management and engineering leaders when reality moves, and make the consequences of each option clear.
Plan AI-enabled product work with the discovery and evaluation loops it genuinely requires, and hold the line on shipping only when quality has been measured against real user scenarios.
Lead the scrum masters and set a consistent standard for planning, backlog health, estimation and demo cadence across squads, without turning agility into ceremony.
Improve how work flows: shorten cycle time, limit work in progress, and reduce the handoffs and wait states that quietly cost more than any individual delay.
Instrument delivery with a small set of trusted metrics, and use them to spot systemic problems rather than to grade teams.
Clear the impediments that a squad genuinely cannot clear on its own, and fix the recurring ones at their source.
Put AI to work inside the delivery function itself, using it for planning support, status synthesis, risk detection and reporting, so the team spends its time on judgment rather than on assembling slides.
Own ship readiness for each product increment: what is in it, who it affects, how it is enabled, and what happens if it needs to be pulled back.
Coordinate the non-engineering half of a launch, working with product management, design, support and market teams on enablement, documentation and communications so users are not surprised by their own product.
Tune the bar to the audience: an internal tool can iterate in the open, while an externally used product carries commitments to customers that shape sequencing and timing.
Partner with Delivery and Operations on the first weeks after launch, then close out heightened support once the product is genuinely stable.
Engage with the architecture and the trade-offs directly: this role is expected to understand why a design choice creates a schedule risk, not merely to record that engineering said so.
Keep a live view of program risk with quantified impact on outcome, scope or date, and bring options to leadership rather than status.
Protect engineering focus by keeping reporting light, decisions fast and meetings few, and by absorbing coordination overhead so squads do not have to.
Program-manage AI work honestly: build and evaluation cycles that are inherently iterative, quality that is measured statistically rather than passed or failed, dependencies on models, data and evaluation sets, and cost and latency characteristics that shift after launch.