Technical Product Manager

At Robots & Pencils, we build meaningful, scalable digital products that solve real business problems. We are looking for a Staff Product Manager who combines deep Generative and Agentic AI fluency with hands-on building ability to own AI product outcomes end-to-end. As a Staff PM, you're accountable for initiative-level outcomes, stakeholder satisfaction, and contributing to R&P's AI product practice. You think in systems, work backwards from the customer problem, and stay relentlessly curious about what's next in AI Enterprise clients want to deploy Agents - moving from a promising demo to a production system that works at scale, meets security and compliance requirements, and delivers measurable business value is hard. This role owns that problem. You'll be part of a GenAI initiative within the AWS ecosystem, building the evals, tools, patterns, and reference architectures that make AI deployment repeatable. The mindset: prove it works, test assumptions early, and document while building. Key Responsibilities Product Strategy & AI Vision Define and drive the product vision, strategy, and roadmap for GenAI solutions - with agentic AI (agent orchestration, tool use, multi-step workflows) as the primary focus - connecting AI capabilities to enterprise business outcomes Translate enterprise problems into structured product requirements; reframe feature requests into outcome-driven priorities with explicit tradeoffs on invest in vs. defer Balance near-term deployment milestones with long-term platform scalability and sustainability Monitor the competitive GenAI landscape and emerging agentic patterns to inform roadmap and technology decisions Discovery & Validation Research how enterprise users interact with AI agents and where they lose trust; frame the riskiest assumptions as testable hypotheses and de-risk them first Design and run experiments - POCs, pilot deployments, scenario-based testing of multi-step workflows, edge cases, and failure recovery - to validate agentic solutions where non-deterministic output makes traditional QA insufficient Distill research, experiments, and competitive intelligence into clear insights that pave the path for a successful product Agent Design, Prototyping & Production Define agent behavior and prototype system prompts and tool schemas; partner with engineering on context management - summarization, working memory, and information flow across multi-step tasks Drive multi-model architecture tradeoffs with engineering - define the quality, cost, and latency targets that determine which model serves each step in the agent workflow Build AI prototypes to validate hypotheses; define human-in-the-loop boundaries and guardrails - when the agent acts autonomously, when it escalates, and how to handle non-deterministic output Establish agent evaluation frameworks - task completion, reasoning quality, tool selection, failure recovery, safety - and partner with engineering on production readiness (observability, drift, responsible AI, prompt versioning) Define success metrics at the agent level - task completion rate, cost per task (not per inference), escalation rate, time to resolution, and customer trust alongside business KPIs Delivery & Execution Own the end-to-end product lifecycle from discovery through phased rollouts; establish the metrics framework (north star, input, guardrail metrics) and report product impact to leadership Manage the product backlog, scope, dependencies, and risks; drive agile ceremonies and produce high-quality PRDs, product briefs, and decision logs Evaluate technology and platform decisions from a product perspective; create deployment playbooks, reference architectures, and knowledge transfer materials so teams sustain solutions independently Use AI to accelerate product work - research, analysis, prototyping, documentation - with judgment on when it needs human oversight; onboard rapidly to new domains and support team members across the initiative Stakeholder Manag…