Product Manager II - Agentic Platform

Toshiba Global Commerce Solutions - External•Durham, NC

6d•Onsite

About The Position

Toshiba Global Commerce Solutions is seeking a Product Manager to help unify and scale our agentic AI platform, the shared foundation that grounds every solution built on it in two decades of accumulated retail expertise. The near-term work is getting the platform's first capabilities launched and consolidating what exists today onto a common core across intelligence and knowledge services, the developer platform, and the fleet of agents that runs on both. It serves teams across Toshiba and the customers and partners who build on it. This is a build-and-ship role: writing specs, working alongside engineering to get capability out, putting it in front of the teams who use it, and bringing back what you learn. You will work closely with platform and engineering leadership on direction, own turning it into well-defined, shipped work, and supply the evidence from real use that shapes what comes next. What Success Looks Like Toshiba's twenty years of retail expertise lives in its products, its architecture, and the judgment behind both. The platform brings that together in one place and puts it to work for anyone who needs it, whether they are building something, answering a question, or delivering for a customer. Success is how far that reaches: engineers shipping faster, expertise available without waiting on the expert, people doing work that used to need a specialist, and customers and partners building on the same foundation.

Requirements

Bachelor’s degree with 10+ years of experience, or equivalent combination.
5-8 years of product management experience, ideally including some exposure to platform, developer-tools, or infrastructure products.
Working familiarity with agentic AI concepts (tool use, retrieval/RAG, multi-agent orchestration, agent evaluation) and genuine interest in building deeper hands-on fluency.
Ability to write clear product requirements and specs, and to prioritize a backlog.
Enough technical fluency to work day-to-day with engineering on APIs and platform architecture, comfortable asking questions and following trade-off discussions without necessarily driving architecture decisions.
A builder's instinct: you use the tools you manage, try things yourself before writing the spec, and can show what you mean rather than only describe it.
Experience supporting an internal-facing product, or a strong interest in internal/platform product work over pure customer-facing feature work.
Comfortable operating with ambiguity and escalating appropriately rather than needing fully defined requirements.

Nice To Haves

Prior exposure to RAG, LLM-based assistants, or agent frameworks in a shipped product.
Familiarity with evaluation, tracing, or observability for LLM and agent systems.
Familiarity with DORA metrics or platform-engineering measurement concepts (deployment frequency, lead time, developer satisfaction/NPS).
Experience instrumenting a shared service and building the reporting people actually use.
Experience with a platform serving both internal teams and external customers: multi-tenancy, self-service onboarding, packaging, or support-model work.
Exposure to usage metering, cost attribution, or internal chargeback for a shared service.
Experience in retail technology or another vertical with SaaS/on-prem hybrid constraints.
Exposure to working with Legal/Compliance on data-rights or AI risk questions.
Experience supporting an internal tool or platform used by other teams rather than end customers.

Responsibilities

Own the shared platform backlog across intelligence and knowledge services, the developer platform, and the agent fleet.
Write clear product requirements and specs, translating business use cases from internal stakeholder teams into actionable work.
Partner day to day with Engineering/Architecture on platform structure, flagging feasibility issues and trade-offs.
Support execution of the governance model for agent build, review, and release: registry entries and ownership, risk tiers, human-in-the-loop review, evaluation thresholds, promotion criteria, and the data access and customer data protections that apply at every step.
Ship the curated workflows, skills, and reference patterns that make the platform faster to use and more reliable in what it produces, along with the onboarding documentation that lets a team get going without the platform team in the room.
Operate with a builder mindset: use the platform yourself, prototype what you are asking for, and drive adoption directly with the teams rather than only through documentation.
Close the loop: gather feedback and usage signals from the teams on the platform, turn friction into concrete backlog items, and bring the evidence that informs the roadmap.
Build the instrumentation and reporting behind the platform's KPIs, so progress is visible to the teams and to leadership: adoption and breadth of use across teams, onboarding time, the share of portal usage that reaches merged and deployed code, developer experience, agent and retrieval quality, and usage cost and attribution.